Power and Personality: An Analysis of Gay Male Intimate Abuse

Monica A. Landolt and Donald G. Dutton

University of British Columbia

Abstract

A study was conducted to investigate the issue of gay male intimate abuse in a sample of 52 couples. The objective of the study was to explore the association between relationship power dynamics and the perpetration of psychological abuse, and to investigate the correspondence between intimate abuse and factors associated with the Abusive Personality (Dutton & Starzomski, 1993; Dutton, 1994a). The majority of participants were middle to upper class and their ethnic/racial breakdown is as follows: White (86%), Native American (7%), Asian (4%), and Other (3%). In terms of power, results indicated that the more frequent form of psychological abuse was significantly higher in relationships characterized by divided power (i.e., partners sharing decision-making authority by the partners each making decisions in different domains). This finding supports the supposition that abuse can occur in relatively egalitarian relationships. An association between intimate abuse and the Abusive Personality was also clearly observed and this personality profile seemed to characterize both members of abusive dyads. Conclusions are drawn regarding the generalizability of intimate abuse across lines of sexual orientation.

Power and Personality: An Analysis of Gay Male Intimate Abuse

As in heterosexual relationships, intimate abuse is frequently reported in gay male relationships. In fact, domestic violence is believed to be the third largest health problem facing gay males today following AIDS and substance abuse (Island & Letellier, 1991), although reliable incidence estimates are not available (Letellier, 1994). The aim of this research was to investigate factors which predict intimate abuse among gay male couples. Because there is a dearth of empirical research on the subject of gay male intimate abuse, the conceptual development of the given study will borrow from a number of related extant literatures. Thus, a broad theoretical overview will be outlined below.

One question posed by the frequency of abuse in gay relationships is whether explanatory models of abuse, developed in heterosexual relationships, would apply to gay male intimate abuse? There is a substantive body of research investigating heterosexual domestic violence which follows from gender-based theories that place the root of assault in the patriarchal of society (e.g., Yllo & Straus, 1990; Smith, 1990). Wife assault is not seen as an anomalous event transpiring only in the most dysfunctional families, but rather is understood as a systemic manifestation of male perceived right of dominance over women (Bograd, 1988; Dobash & Dobash, 1979; Walker, 1989). Indeed, in their nationally representative sample of 2,143 couples, Coleman and Straus (1986) found that male-dominant couples had the highest rate of conflict and that there was a concomitant high violence risk in those couples characterized by such a power-imbalance.

However, the data are highly contradictory and at best, indicate that there is not a simple linear relationship between domestic violence and patriarchy. For example, contrary to the patriarchy perspective, female-dominant couples were also found to have high rates of conflict (Coleman & Straus, 1986). Gelles and Straus (1988) reported that while the rate of husband-to-wife violence was 113 per 1000, the wife-to-husband violence rate was 121 per 1000 couples. Hamberger and Potente (1994), however, interpret wife-to-husband violence rates in terms of retaliation or self-defense which concurs with the patriarchal perspective. Similarly, Saunders (1988) reported that when women do use violence, it is more likely to be against a violent partner than a non-violent partner. Conversely, Stets and Straus (1990) reported that in violent couples, the pattern of female use of severe violence and male use of minor violence was more prevalent than the reverse pattern (male use of severe violence/female use of minor violence) across samples of dating, cohabiting and married couples. Finally, data clearly indicate that domestic violence outcomes generated by male-to-female assault are more serious and damaging than female-to-male assault (e.g., Berk, Berk, Loseke, & Rauma, 1981; Browne, 1992; Stets & Straus,1990).

Obviously there are a number of contradictions apparent across the various empirical studies. Beyond this limitation, a major problem implicit in the patriarchal perspective on wife abuse is that it only recognizes heterosexual male violence, thus it fails to address the conflict dynamic in same-sex relationships (Letellier, 1994). Some researchers do acknowledge the patriarchal perspective’s failure to accommodate the gay victim’s experience (e.g., Martin, 1976). Their reconciliation strategy has been to impose a heterosexual gender-based model onto gay relationships thereby assuming that within such couples, there is an assigned masculine "husband" (who is most probably the perpetrator of abuse) and a corresponding "wife" or effeminate gay male (who is probably the victim of abuse). However, in an unpublished study of abuse in gay male and lesbian couples, Kelly and Warshafsky (1987) found that neither a strong masculine, feminine, nor an androgynous gender role identity was related to the use of aggression in relationship conflicts. Aggression was found to be most prevalent among those individuals who had an undifferentiated gender role identity (e.g., reported having both a low masculine and feminine gender role identity).

Gay male intimate abuse would be more consonant with a patriarchal perspective if it could be framed in terms of entrenched gender roles. However, much research indicates that gay male effeminacy is more of a stereotype than a reality. It is accurate to say that effeminacy may be more characteristic of gay men than heterosexual men (Saghir & Robins, 1973; Whitam, 1977), but the majority of gay men are not effeminate, and do not express any interest in being female (Harry, 1982). Moreover, as Harry (1984) asserts, effeminate gay men are individuals who are generally high in self-esteem, dominance, and competitiveness. Given this finding that most gay men are not effeminate, ergo, most gay couples do not include a "functional" wife: what is the most prevalent relationship pattern in gay male relationships? More specifically, how frequently does heterosexual role playing occur in gay male relationships?

There is general consensus in the literature that male/female role playing in gay male relationships is the exception rather than the norm (e.g., Bell & Weinberg, 1978; Peplau, 1993). Harry (1984) contends that homogamous selection is the key defining characteristic of most gay male relationships; individuals are attracted to partners who have a similar degree of masculinity or femininity. Thus, butch/butch or femme/femme relationships are more likely than the supposed butch/femme pattern. Kelly and Warshafsky (1987) report findings that generally concur with this homogamous selection hypothesis. In their sample of gay and lesbian couples, cross gender pairing (couples including both a masculine and a feminine partner) was very rare. While 36% of their subjects shared a similar gender role identity with their partner, in most cases crossing was between a male role, or to a lesser extent a female role (26% of the gay male sample were effeminate) with an androgynous or undifferentiated gender role. Finally, Harry (1984) asserts that "best friends" rather than husband and wife is the best way to characterize the majority of gay male relationships.

Given the assumption that the primary relationship pattern of coupled gay males is egalitarian-friendship, and the reported finding that masculine/feminine role playing is not typical of such relationships; one might conclude (albeit prematurely) that intimate abuse is incompatible with the image of the gay male relationship. Indeed, it could be argued that, because both partners share a similar socialization and perspective (demonstrated by homogamous selection), there might be a concomitant harmony in the relationship that is not experienced by heterosexual couples (Peplau & Gordon, 1983). However, the egalitarian nature of gay male relationships can be contrasted with Symons’ (1980) assertion that, "in homosexuality, we see male and female sexuality in its purest, uncompromised form". Gay men are biologically male and culturally conditioned as such (Reece & Segrist, 1981). Thus, we could extend Symons’ proposition and argue that because gay male relationships necessarily involve two men, there may actually be a heightened level of dominance and control observed among these couples.

Returning to the heterosexual domestic violence literature, a promising line of research investigating this form of abuse revealed that 80% to 90% of men in treatment have diagnosable personality disorders (Hamberger & Hastings, 1986; 1991; Hart, Dutton, & Newlove, 1993). This rate is striking given that the rate of personality pathology found in the general population is only 15-20% (Kernberg, 1977; Zimmerman & Coryell, 1989). Island and Letellier (1991) also implicate psychopathology as a causal factor in gay male intimate abuse. However, their presumption is not based on empirical research, but rather on clinical and anecdotal evidence. Farley (1996) conducted psychological assessments of 119 gay men in treatment for perpetration of intimate abuse. He found that 87% of his sample reported having high levels of previous mental health/psychiatric difficulties.

A personality profile of psychologically and physically abusive heterosexual men has been generated by Dutton (1994a) and further validated by Dutton, Starzomski, & Ryan (1996). A central feature of this profile includes a form of personality disorder referred to as Borderline Personality Organization (BPO; Oldham et al., 1985). Additional features of the profile include attachment factors and experiential factors in the perpetrator’s family of origin. This constellation has been described as an "Abusive Personality" (see Dutton, 1994a; Dutton, Starzomski, & Ryan, 1996) since personality variables account for an impressive amount of the variance in partners’ reports of psychological abuse, and to a lesser extent, physical abuse. The Abusive Personality was also predictive of psychological abuse in a group of heterosexual males not formally identified as abusive (Dutton, 1994a).

BPO, the central component in the Dutton constellation can be characterized by: 1) a proclivity for intense, unstable interpersonal relationships expressed in intermittent undermining of the significant other, manipulation and masked dependency; 2) an unstable sense of self with an intolerance of being alone and abandonment anxiety; and 3) intense anger, demandingness and impulsivity, usually tied to substance abuse or promiscuity (Gunderson, 1984). BPO is understood in terms of a continuum with Borderline Personality Disorder (BPD) at one extreme pole.

Silverstein (1988) suggests that BPD (the more extreme form of BPO) may be more prevalent among gay men. This may be due to the fact that identity development can be impaired by self (i.e., internalized homophobia), and other’s negative attitudes regarding homosexuality (i.e., external homophobia) (Harry, 1984; 1989). Indeed, many gay males report childhood experiences that differ from those of heterosexual males on key dimensions such as effeminacy (Bell, Weinberg, & Hammersmith, 1981; Saghir & Robins, 1973), and that defeminization only occurs in late adolescence primarily due to social pressure (Harry, 1982). Therefore, it is possible that this childhood environment could contribute to a more precarious sense of self and hence, a greater proclivity for Borderline Personality Organization.

Abandonment anxiety is one of the constituent elements of BPO. More precisely, the form of abandonment anxiety expressed by fearful attachment (and to a lesser degree, preoccupied attachment) has been closely associated with the Abusive Personality. In particular, Dutton, Saunders, Starzomski, and Bartholomew (1994) found that fearful and preoccupied attachment were prevalent and correlated significantly with scores on BPO, anger, and jealousy in a group of heterosexual men convicted for wife assault.

Fearful and preoccupied attachment refer to types of response styles experienced by persons when they are in intimate relationships and are two of the four categories in Bartholomew’s typology of adult attachment (Bartholomew, 1990; Bartholomew & Horowitz, 1991). Bartholomew’s model is rooted in the two dimensions of negative/positive self-regard and negative/positive other-regard. Specifically, fearful individuals have a negative regard for both self and other and while they desire social contact and intimacy, they experience pervasive distrust and fear of rejection. Their relationships are characterized by a high degree of anxiety and relationship anger. Preoccupied attachment describes those individuals who are overly dependent, while their regard for the significant other is high, self-regard is generally low. Thus, these individuals have a strong feeling of unworthiness expressed by an incessant need to please others, as well as make constant demands on others.

In addition to fearful attachment and preoccupied attachment, Bartholomew’s model also includes secure and dismissing attachment styles. Secure attachment is characterized by a general comfort with intimacy and autonomy. These individuals have a positive regard for both themselves and their significant other and, therefore, typically have secure and fulfilling adult relationships. Finally, individuals with a dismissing style have a positive self-image but a low regard for others; a premium is placed on independence and the result of this is the general avoidance of close relationships. When such relationships are established and rejection results, these individuals distance themselves and devalue the importance or worth of the relationship.

Dutton, Starzomski, and Ryan (1996) have also found that recollections of negative childhood experiences (including experiencing physical abuse) are highly associated with the Abusive Personality. In general, research has indicated that there is a connection between early trauma and the development of BPO (van der Kolk, 1987; Herman, Perry, & van der Kolk, 1989). Specifically, Dutton et al. (1996) have found that paternal rejection and to a lesser degree absence of maternal warmth, accounted for a significant portion of the variance in the BPO scores of male batterers. Thus, Dutton tentatively concluded that in the heterosexual example, parental mistreatment could very well be an antecedent to the development of an abusive personality.

This connection between negative childhood events and subsequent problems in intimate relationships also seems to be relevant for gay men. Harry (1989) has found that gay men, as children, were significantly less attached to their fathers than heterosexual men and that as adolescents, gay men were more likely than heterosexual men to have suffered physical abuse perpetrated by their parents. McWhirter and Mattison (1984) report that when one or both partners in a gay male couple express childhood memories of parental alienation, there is an observed increased difficulty in conflict resolution in the current relationship. Similarly, Farley (1996) reports that in his sample of 119 gay male perpetrators of intimate abuse, 93% of the men had experienced childhood physical abuse and 67% had experienced childhood sexual abuse. Thus, the impact of recollecting negative childhood experiences may be particularly salient in the case of gay male intimate abuse, increasing the likelihood that there is a significant relationship between the Abusive Personality and perpetration of such abuse.

There were two principal goals of the present study. The first was to examine the issue of power arrangements observed in gay male couples. The extant literature seem to indicate that most gay relationships are characteristically egalitarian, we wanted to confirm this observation, as well as investigate the association between relationship power and the perpetration of psychological abuse. In particular, we asked the question, if psychological abuse is present in the gay male sample, is it more likely to occur in relationships characterized by inequality in decision making? The second goal was to investigate the correspondence between factors associated with the Abusive Personality (e.g., BPO, fearful attachment, preoccupied attachment, and poor child/parent relationships) and the presence of intimate abuse.

Method

Fifty-two gay male couples participated in the study. Participants provided information regarding age, education (last year completed), occupation, ethnic background, relationship status, length of time in current relationship, and income (represented by 7 intervals ranging from less than $10,000 to over $60,000). The demographic profile of the men in the (entire) sample is as follows: average age: 34 years (range 20-64), average level of education: 2 years college or university. The ethnic background of the sample was: White (86%), followed by Native American (7%), Asian (4%), and Other (3%). The mean length of relationship was five and a half years (range six months to 31 years). The modal joint annual income of couples was $60,000. The majority of participants lived in a large urban center while a small minority lived in suburban locations.

Relationship power. The Marital Power Scale (MPS; Coleman & Straus, 1986) is a modified version of the Decision Power Index developed by Blood and Wolfe (1960) and was further revised for use with a gay sample. This instrument asks respondents "Who has the final say" in decision making about the following seven issues: buying a car, whether to buy life insurance, what house or apartment to take, whether a partner should go to work or quit work, how much money to spend on food, what job either partner should take, and where to go on vacation. The Coleman and Straus (1986) findings are based on a nationally representative sample of 2,143 heterosexual couples (either married or cohabiting). Participants were classified into one of the four relationship power categories: egalitarian, partner dominated, self dominated, or divided power, following the method used by Coleman and Straus (1986). The difference between the egalitarian and divided power types is that the former are equal in the sense of making most decisions jointly, whereas the latter are equal in the sense of dividing responsibility for decisions, with each individual in the couple having a final say for different decisions (Coleman & Straus, 1986).

Borderline Personality Organization. The Self-Report Instrument for Borderline Personality Organization (Oldham et al., 1985) is a 30-item instrument derived through factor analysis of a 130-item questionnaire designed by the authors. The scale retains items with the strongest factor loadings for each of the three subscales of identity diffusion, primitive defenses and reality testing. For the purpose of this study, a general measure of BPO was obtained by summing together the three sub-scales. The Cronbach alpha for the overall BPO measure was .91. The BPO self-report instrument does not assess abusiveness or aggression. Hence, any associations with other scales reported below are not attributable to item overlap.

The first subscale, identity diffusion, measures a poorly integrated sense of self or of significant others. Identity diffusion is assessed by ascertaining difficulties in describing one's own personality or the personalities of others, uncertainty about career or goals, contradictory behaviors, and instability in intimate relationships. The second subscale measures primitive defenses including: splitting, idealization, devaluation, omnipotence, denial, projection and projective identification. The third subscale, reality testing, measures external versus internal origins of perceptions, evaluation of one’s own behavior in terms of social criteria of reality, differentiation of self from non-self, internal reality testing, and the cognitive process of reality testing.

Anger. The Multidimensional Anger Inventory (MAI; Siegel, 1986) is a 38-item self-report scale assessing the following dimensions of anger response: frequency, duration, magnitude, mode of expression, hostile outlook, and range of anger-eliciting situations. The scale was validated by correlation with other, conceptually similar anger inventories. . The Cronbach alpha for this scale was .91.

Attachment. Relationship Scales Questionnaire (RSQ; Griffin & Bartholomew, 1994) is a 30-item self-report measure with items drawn from Hazan and Shaver's (1987) attachment measure, Bartholomew and Horowitz's (1991) Relationship Questionnaire, as well as items from Collins and Read's (1990) Adult Attachment Scale. Measures of each of the four attachment patterns (secure, fearful, preoccupied, dismissing) identified by Bartholomew and Horowitz (1991) were created by summing four or five items from the corresponding prototypic descriptions. Each participant in this sample received a continuous rating for each attachment pattern. Cronbach alphas for the two theoretically relevant sub-scales (fearful and preoccupied attachment) were .59 and .35 respectively. Given the low internal consistency, these two subscales are conservative measures and, therefore, any associated, significant findings should be considered quite impressive.

Recollections of early childrearing. Egna Minnen Betraffande Uppfostran (EMBU; Perris, Jacobsson, Lindstrom, von Knorring, & Perris, 1980) scale was used to provide a quantitative measure of the respondents' memories of their upbringing. The EMBU was originally developed in Sweden and has been translated and widely used with English speaking samples (Gerslma, Emmelkamp, & Arrindell, 1990). It is an 80-item scale that assesses memories of parental rearing behavior. The psychometric properties of the English version were developed by Ross, Campbell, and Clayter (1982). The English version has 14 subscales, scored separately for mother and father. For purposes of this study, only subscales assessing recollections of maternal and rejection were assessed, comprising 43-items. The rejection subscale included items pertaining to both physical and psychological abuse. An example of a physical abuse item from the rejection subscale is, "My parents beat me for no reason." An example of a psychological abuse item from the rejection subscale is, "My parents treated me in such a way that I felt ashamed." The Cronbach alphas for the paternal and maternal rejection sub-scales were .96 and .93 respectively.

Two abuse measures were used in this study. However, given that the gay sample is drawn from the general population and has not been formally identified as abusive, psychological abuse, rather than physical abuse, is used as the primary dependent measure.

There are a number of theoretical reasons for using a psychological abuse measure over a physical abuse measure as the primary dependent variable. In the heterosexual domestic violence literature, it has been well documented that psychological abuse is a precursor and often accompanies physical abuse (Tolman, 1989; Walker, 1979). Similarly, self-reports by battered women suggest that the effects of psychological abuse on self-esteem and recovery are more prolonged and emotionally debilitating than the immediate impact of physical abuse (Dobash & Dobash, 1979; Walker, 1984). Finally, it has been observed that while many men are able to stop physical abuse after arrest and treatment, they often continue or even increase the use of psychological abuse as a way of maintaining control (Tolman, 1989).

Physical abuse. The Conflict Tactics Scale (CTS; Straus, 1979) is a standardized scale designed to measure the frequency and intensity of 19 tactics used in dyads to resolve conflict. The scale includes rational tactics, withdrawal, and a variety of verbally, emotionally and physically abusive strategies. However, only the total physical abuse sub-scale was used in the analyses. The total physical abuse subscale includes eight items ranging from throwing something at an individual to using a knife or gun on an individual. Respondents report both their own use of these tactics, as well as their partners’ use of the tactics on a scale ranging from 0 (never) to 6 (over 20 times). The Cronbach alpha for the total physical abuse sub-scale was .87.

Emotional abuse. Psychological Maltreatment Inventory (PMI; Kasian & Painter 1992) was used to assess emotional abuse. The PMI is based on Tolman’s (1989) Psychological Maltreatment of Women Inventory (PMWI) but the PMI can be used with both men and women. Where relevant, the wording was revised to accommodate a gay male sample. The PMI contains 58-items rated from "never" (0) to "more than twenty times" (6) that cover forms of emotional and verbal abuse which have occurred during the last year. The PMI has five psychological maltreatment subscales: jealousy, isolation and emotional control, withdrawal, verbal abuse, and undermining of partner’s self esteem which were aggregated to form a general measure of psychological abuse. The Cronbach alpha for the general measure of psychological abuse was .95. Participants completed a PMI based on their partner’s behavior.

Couples responded to advertisements placed in two local gay and lesbian newspapers. The initial contact was made via the telephone, at which point participants were asked the status of their sexual orientation (as well as their partner’s) and it was confirmed that they had been together for at least six months. Both partners were mailed two identical questionnaire packages along with two self-addressed, stamped envelopes.

In order to ensure honesty in responding, participants were asked to fill out questionnaires separately, without partner consultation. Identification numbers were assigned to each individual to ensure anonymity. When completed questionnaire packages were returned by both members of the couple, each participant was mailed a $30 payment fee. Measures to ensure confidentiality were employed; participants were informed that payment forms which included addresses would be kept separate from the body of the questionnaire and would be destroyed at the point of payment. In all cases where there was some uncertainty regarding a participant’s response to a questionnaire item (e.g., in the case of missing data), follow-up telephone calls were made by the first author. This procedure was employed to ensure a consistent quality of data. The measures discussed (listed above) were part of a larger assessment battery.

Participants responded enthusiastically to the advertisement. Approximately 75 calls were received overall but only the first 55 callers who met the requirements of the study were accepted for participation. The high degree of interest in the study was most probably due to the fact that members of the gay community strongly desire to participate in research relevant to gay male well-being and mental health (P. Lettelier, personal communication, February 1997). Mailing out questionnaires, as well as including stamped, self-addressed envelopes reduced concerns regarding public disclosure of sexual orientation, and therefore, more likely contributed to a representative sample. No couple who showed initial interest (e.g., made the telephone call) and qualified for the study, declined to participate after hearing details of the study. The response rate for completed and returned questionnaires was .94.

A debriefing form was included at the end of the questionnaire package which encouraged participants to call the laboratory for a referral to a qualified counselor if they were disturbed by some of the sensitive issues raised in the questionnaire. Only one couple called with concerns regarding the level of discord in their relationship and an appointment with the counselor was arranged.

Results

These data came from interdependent couples. As such, standard statistical techniques that assume independence of observations cannot be used (Gonzalez & Griffin, 1996; Kenny, 1988). To control for this, we examined interdependence in couples using intraclass correlations. When intraclass correlations were significant, then special techniques that corrected for degree of interdependence were used (Gonzalez & Griffin, 1996)

Forty percent of the sample reported that at least one member of the couple perpetrated one or more violent acts in the last year (21/52; based on partners’ reports on the Conflict Tactic Scale total physical violence subscale). The correlation between self-report of self-perpetrated (actor) violence and partner report of actor-perpetrated violence is r =.72, p < .001, verifying the accuracy of the self-report measures.

The first research question investigated the relationship between psychological abuse and decision-making power. Although participants could be placed in one of four categories based on their MPS (power) scores, in actuality, participants reported membership in only three of the categories which were: egalitarian (n=59), self-dominated (n=8), and divided power (n=30). Power scores were missing for seven participants; thus the total number in this analysis was 97.

Preliminary analyses were conducted to assess agreement between individuals within each couple regarding the self-perceived relationship dynamic. A pairwise intraclass correlation was generated for perceptions of power and was found to be non-significant r = -.08 (z = .-61). This indicated that within each couple, individual members did not necessarily share similar perceptions of the relationship’s power dynamic.

Comparisons were conducted to assess whether psychological and/or physical abuse was higher in couples who showed discongruent perceptions of power. Power-discongruence was based on discordant categorical data (i.e., members of the dyad were placed in a different power category according to Coleman & Straus’ (1985) methodology for categorizing self-report power scores). For each of the two dependent variables (psychological abuse and physical abuse), dyad mean scores were calculated by averaging across the two individual member’s abuse scores. The mean psychological abuse score for power-congruent couples was 67.07 (SD = 19.22), the mean psychological abuse score for power-discongruent couples was 67.66 (SD = 20.52). A t test revealed that these two means were not significantly different. In terms of physical abuse, the mean level of abuse for power-congruent couples was 1.43 (SD = 4.12) and the mean level of abuse for power-discongruent couples was 2.28 (SD = 4.91). Again, a t test revealed that there was no significant difference between the two means. Thus, while members of a couple may disagree on self-perceived relationship power dynamics, this did not seem to impact significantly on corresponding levels of psychological or physical abuse.

Tests were then conducted to see if psychological abuse varied as a function of power category. The dyad was again the unit of study, although power and abuse data from each member of the couple was used in the overall analysis. To this end, individuals within each couple were assigned to one of two categories based on their psychological abuse scores (PMI; Kasian & Painter, 1992). The individual with the higher PMI score was coded as the "abuser" and the individual with the lower score was coded as the "victim". In three cases, both members of the dyad had identical PMI scores, therefore one member of the couple was randomly assigned to the abuser category. A significant t test revealed as expected, that psychological abuse perpetrated by the abuser (M = 75.39, SD = 24.68) was indeed higher than that perpetrated by the victim (M = 61.62, SD = 16.62), t (51) = 6.54, p = 000.

The data were then analyzed with a 3 X 3 X 2 repeated measure analysis of variance. Abusers’ Perception of Power (three levels: egalitarian, self-dominant, and divided-power) and Victims’ Perception of Power (three levels: egalitarian, self-dominant and divided-power) were between-dyad factors. Abuse Perpetrator (two levels: one indicating abuse done by the victim, and the other referring to abuse done by the abuser), functioned as the with-in dyad, repeated measure.

A significant Victims’ Perception of Power by Abuse Perpetration interaction was found, F (2, 38) = 4.00, p =.027, and is depicted in Figure 1. However, neither the Abusers’ Perception of Power by Abuse Perpetration interaction was significant, nor was the Abusers’ Perception of Power main effect significant. Decomposition of the significant interaction revealed that abuse perpetrated by the victim varied significantly as a function of victims’ perception of power. Specifically, psychological abuse (perpetrated by the victim) was highest in self-dominant couples as compared to egalitarian couples, t (38) = 2.77, p = .008, and there was no significant difference between self-dominant and divided-power couples. Abuse perpetrated by the abuser (the more frequent case of psychological abuse), also varied significantly as a function of victims’ perception of power. However, at this level, further analyses revealed that psychological abuse (perpetrated by the abuser) was significantly higher in divided power couples than in egalitarian couples, t (38) = -2.24, p =.03, while no significant difference was observed between divided power and self-dominant couples.

BPO was correlated with the other constituent elements of the Abusive Personality. A significant positive relationship was found between BPO and anger, fearful attachment, preoccupied attachment, paternal rejection, and maternal rejection. The clear and consistent relationship between BPO and the other independent variables seems to indicate that BPO is a central feature in the Abusive Personality. Overall correlations (between personality measures) and cross-intraclass correlations (between personality and abuse measures) depicting the theoretical and statistical centrality of BPO in this constellation are presented in Figure II.

The association between the Abusive Personality and intimate abuse was then assessed by correlating the various personality and abuse measures. Results of these analyses are presented in Table I. As indicated, partner reports of receiving psychological abuse were significantly correlated with self (actor)-reports of BPO, anger, fearful and preoccupied attachment, recollections of paternal rejection, recollections of maternal rejection and negatively correlated with secure attachment.

Table I also reveals that partner reports of receiving physical abuse were significantly correlated with self (actor)-reports of BPO, fearful and preoccupied attachment, and recollections of maternal rejection but not paternal rejection. There was no significant association between physical abuse and secure attachment (although results were in the intuitive direction).

Given that comparable data was collected from each member of the dyad, we can ask the question, do personality variables in the actor better predict actors’ perpetration of psychological and physical abuse than do partner personality variables? More specifically, using the method suggested by Gonzalez and Griffin (1996), we can determine the relative contribution of the Abusive Personality in the actor, versus the Abusive Personality in the victim, in terms of the actor perpetrating psychological and physical abuse.

To test this question, separate, simultaneous multiple regressions were conducted for each constituent element of the Abusive Personality (including BPO, anger, fearful attachment, preoccupied attachment, maternal rejection and paternal rejection). Psychological and physical abuse perpetrated by the actor were predicted from both partner and actor personality variables. Results of these analyses are depicted in Table II. As indicated, for psychological abuse, both actor and partner effects were significant. In the case of physical abuse, both actor and partner effects were significant for BPO, fearful, and preoccupied attachment; neither actor nor partner effects were significant for anger and paternal rejection, and only actor effects were significant for maternal rejection.

Pairwise intraclass correlations were generated for the two abuse variables, as well as for all personality variables. Significance was observed for psychological abuse (r = .57, z = 4.11, p < .001), and physical abuse (r = .57, z = 4.12, p < .001). These results indicate that when one member of the dyad was abusive, it was also more likely that his partner was abusive. Significance was also observed for two personality variables; BPO (r = .46, z = 3.34, p < .001), and paternal rejection (r = .39, z = 2.84, p < .01). These latter results suggest that individuals with some elements of the abusive personality (BPO and paternal rejection) were paired with similar partners. Given the above findings, we could explore the possibility that dyad levels of abuse, and dyad levels of abusive personality, might differ as a function of relationship abuse dynamics (mutually abusive versus uni-directionally abusive).

Thus, we assessed whether average levels of abuse were higher in mutually combative relationships (measured in terms of physical abuse) than in uni-directionally violent relationships. Based on partner-reports of physical abuse, abusive couples were classified into one of two abuse categories (mutually abusive; n = 10, or uni-directionally abusive; n = 11). Dyad mean scores were calculated for the two dependent variables (psychological and physical) by averaging the two members abuse scores. A t test comparing mean levels of psychological abuse in mutually abusive relationships (M= 95.30, SD= 17.14) versus uni-directionally abusive relationships (M= 69.54, SD = 15.4) revealed that there were significant differences, t (19) = 3.63, p=.002. Similarly, a t test comparing mean levels of physical abuse in mutually abusive relationships (M= 6.95, SD = 7.31) versus uni-directionally abusive relationships (M= 1.73, SD = 1.72) was also significant, t (9.91) = 2.20, p = .05. These latter two findings indicate that both psychological and physical abuse were higher in those relationships where both individuals used physically abusive tactics.

A number of t tests were also conducted to see if elements of the abusive personality differed as a function of the relationship abuse dynamic (mutually abusive versus uni-directionally abusive). Dyad mean scores were calculated for each of the constituent elements of the Abusive Personality (BPO, anger, fearful and preoccupied attachment, paternal and maternal rejection) by averaging the two members scores on each variable. A significant difference was observed in the case of BPO, t (19) = 3.36, p = .003, preoccupied attachment , t (19) = 2.57, p = .02, and for paternal rejection, t (9.77) = 2.60, p = .03. More specifically, these results indicated that dyad levels of BPO, preoccupied attachment, and paternal rejection were higher in mutually abusive relationships as compared to uni-directionally abusive relationships. Observed means (standard deviations in brackets) for BPO in mutually abusive relationships compared to uni-directionally abusive relationships were 72.90 (9.98) and 57.59 (10.80) respectively. Observed means (standard deviations in brackets) for preoccupied attachment in mutually abusive relationships compared to uni-directionally abusive relationships were 13.20 (1.99) and 11.27 (1.42) respectively. Observed means (standard deviations in brackets) for paternal rejection in mutually abusive relationships compared to uni-directionally abusive relationships were 51.05 (18.98) and 35.09 (4.12) respectively.

Discussion

The first research question examined the issue of decision-making power in gay male couples. Eight percent of the sample reported an imbalance in power (8/97), compared to the heterosexual rate of 16.9% (Coleman & Straus, 1986). Notably, the majority of couples in the present study followed the egalitarian-friendship relationship pattern suggested by Harry (1984). Within couples, agreement regarding perceived relationship dynamic was not typical (as evidenced by the non-significant pairwise intraclass correlation). However, neither psychological nor physical abuse rates were higher in couples who disagreed on the perceived power dynamic in their relationship. This last finding runs contrary to comparable heterosexual results. Specifically, Coleman and Straus (1986) found in the latter case, when consensus regarding power-sharing dynamics was low, conflict greatly increased.

In regard to the relationship between self-ratings of power and psychological abuse, inconsistent findings emerged. No significant relationship between power ratings and abuse was observed when looking at the abusers’ perceptions of power, but a significant interaction was revealed in the case of victims’ perceptions of power. Specifically, in the case of abuse perpetrated by the victim (by definition, the less frequent abuse perpetrator), psychological abuse seemed to be highest in self-dominant couples and lowest in egalitarian couples. This finding is consonant with Coleman and Straus (1986) finding that heterosexual abuse was highest in couples where there was an imbalance in power. However, in the case of the abuser (the more frequent abuse perpetrator), the highest rate of abuse was found in divided power relationships. This latter finding is in direct contrast with the Coleman and Straus (1986) finding of lower abuse rates in equal power relationships (i.e., egalitarian and divided-power relationships).

Coleman and Straus’ (1986) assumption that divided-power relationships are basically egalitarian (although individuals within a dyad express power in different domains), must be accepted with some degree of caution. It may be that members of a couple each have their own domains of power, but this does not guarantee that each domain is equally important. Of course, differential weighting of power domains is probably less common in gay male relationships, than in heterosexual relationships (where gender often defines relationship roles), making our observed finding regarding divided-power and gay male abuse even more surprising.

It is not unusual for inconsistent findings to emerge in a study related to power. The construct of interpersonal power has, historically, been associated with a lack of consensus regarding both definition and operationalization (Olson & Cromwell, 1975; Gray-Little & Burk, 1983). Although, to date, a consensually recognized, superior measure has not been developed (Huston 1983). Moreover, in the present study, the measure of power used (i.e., the MPS) was not exceedingly comprehensive as it included only seven items.

Methodological concerns aside, Merrill (1996) has identified circumstances where gay male abuse can occur in ostensibly equal (e.g., divided-power) relationships. He asserts that opportunity for abuse exists not only when recipients have less social power (or more social power, in our less frequent case of victim perpetrated abuse), but also when they have roughly equal power; that is, if the potential victim is perceived as being unwilling or unlikely to report abuse, or if it is perceived that reporting of abuse will have no effect. Thus, perpetration of abuse may have more to do with recipients’ failure to recognize abuse, or their lack of knowledge regarding the availability of social support options, rather than relationship power dynamics. Clearly, this power and abuse relationship needs to be explored further using more complex indices of power, as well as outcome measures that explore willingness to report abuse and perceptions of social support.

The second goal of this study was to examine the relationship between intimate abuse and the Abusive Personality. Results indicate that the Abusive Personality is clearly present in this gay male sample. Perpetration of psychological and physical abuse (as measured by partner reports of receiving abuse) is correlated with the constituent elements of the Abusive Personality: BPO, anger, fearful and preoccupied attachment and recollection of poor child/parent relationships (although anger and recollection of paternal rejection are not significantly associated with physical abuse).

Interestingly, significant, positive, pairwise intraclass correlations were found for BPO (a central feature of the Abusive Personality) and paternal rejection. The BPO finding is particularly compelling as it suggests that those individuals who show manifestations of the abusive personality, have partners who exhibit similar personality tendencies. The finding that gay men who recollect paternal rejection tend to be coupled with partners who report similar experiences is also important given McWhirter and Mattison’s (1984) assertion that an inability to deal effectively with tension and conflict in (adult) gay intimate relationships is associated with memories of parental alienation during childhood. Moreover, further evidence to support McWhirter and Mattison’s (1984) claim can be found in the fact that dyad levels of paternal rejection were higher in mutually abusive gay male relationships (as compared to uni-directionally abusive relationships).

Results of the separate multiple regressions revealed that for the most part, both partner and actor abusive personality traits significantly predicted actor’s perpetration of psychological and physical abuse, except in the case of maternal rejection where only actor effects were significant. At first glance, this finding may seem highly controversial as it could be interpreted in a way that blames the victim for his receiving psychological and physical abuse. However, recall that significant, positive, pairwise intraclass correlations were observed for both psychological and physical abuse indicating that when one member of the dyad is abusive it is likely that his partner is also abusive. Moreover, it seems to be the case that when abuse occurs, both members of the couple are likely to exhibit Abusive Personality tendencies. This conclusion is further substantiated by the fact that levels of abuse, and abusive personality tendencies are significantly higher in mutually abusive relationships as compared to uni-directionally abusive relationships. Thus, it can be established that in the case of gay male intimate abuse, the Abusive Personality is particularly relevant as it seems to pertain to both members of the dyad.

Limitations

The procedure employed to administer questionnaires (i.e., via the mail) was both a strength and a weakness of this study. While it certainly allowed participants to respond in an anonymous manner, a drawback of this method is that no researcher was present to observe the actual completion of the questionnaires. Efforts were made to safeguard the quality and integrity of the responses. For example, after all data had been collected, a selection of participants were re-contacted and asked whether they had consulted their partners during the completion task. All of these participants indicated that indeed, they had not consulted their partners. Nonetheless, because participants completed questionnaires at home, one can never be certain that they were done so separately and independently. Replication of the study in a laboratory setting may further confirm the authenticity of responses.

The demographic profile in this study was overwhelmingly that of young, middle to high economic status, white men. The underrepresentation of older, blue collar, less educated, and racial/ethnic minority gay men is not specific to this study, but is a common complaint of most researchers in the field of gay male studies (e.g., Bell & Weinberg, 1978; Harry, 1984). The failure to access a more representative sample may very well be due to the recruitment method employed in this study. Harry (1990) conducted a national probability sample of men and found that the profile of gay men not formally associated with the "gay world" to be very different from the profile of gay men recruited in non-probability samples (e.g., as in this study). In the Harry (1990) study, many of the gay men were married, members of minorities, less educated, older in age, and small town residents. A self-selected sample, willing and interested in research, may be an inevitable outcome of securing participants through advertisements in newspapers. Although attempts were made to reduce bias among the volunteer sample used in the present study (see Rosenthal & Rosnow, 1975), its lack of representativeness does hinder our ability to generalize the observed findings to the larger gay male population.

Some may also challenge the perfectness of fit between the measure of power used in the present study (MPS) and gay male relationship roles. While it is true that a multi-dimensional measure of power would be more ecologically valid, the MPS has long been used to investigate power roles in the domestic violence literature and close relationship literature. The measure of power used here allowed us to compare results with extant heterosexual findings, which, at this point, is still necessary given the general dearth of research on gay male intimate abuse. Thus, utilization of a historically grounded measure seems warranted. Although, further studies investigating the relationship between additional forms of power (e.g., power observed in verbal and behavioral interactions or trait measures of power) and gay male intimate abuse would certainly contribute to the nascent literature.

General Conclusions

One conclusion that can be drawn from this investigation is that some dynamics underlying intimate abuse are not specific to male-female relationships but transcend the boundaries of sexual orientation. In particular, this conclusion is based on the correspondence between the findings of this study and the finding that an Abusive Personality is strongly related to perpetration of psychological and physical abuse in samples of heterosexual male batters as well as heterosexual men not formally identified as abusive (Dutton, 1994a; Dutton & Starzomski, 1993). More specifically, the centrality of BPO in the Abusive Personality appears to be consistent across both heterosexual and gay male samples.

Bearing in mind the above stated contribution of personality factors, we can further conclude that intimate abuse does not appear to be a function of male-female gender typed relationship models because of the following lines of evidence: 1) masculinity in gay men is not related to abuse (Kelly & Warshafsky, 1987); 2) most gay men are not effeminate (Bell et al., 1981; Harry, 1982), nor are their relationships divided according to male/female gender roles (Harry, 1984) thus allowing us to conclude that rarely is there a "functional wife" in these relationships; and 3) In the present study, we found that when abuse does occur, it is not necessarily associated with power dominance, but with divided-power, as opposed to the most egalitarian power dynamic. Indeed, our results indicate that abuse can occur in roughly equal relationships.

Therefore, it can be cautiously stated that a direct cause and effect relationship between intimate abuse (at the very least gay male intimate abuse) and societal patriarchy is unlikely. Rather, this form of abuse is likely to stem from personality variables organic to perpetrators (e.g., the Abusive Personality). This is not meant to imply that social factors external to the intimate dyad do not have an effect on the internal workings of the relationship. In fact, it may be that individuals with a propensity for abusiveness actually seek out elements of the ambient culture to justify their actions in an effort to reduce their feelings of guilt and remorse (Dutton, 1994b). Patriarchy should be seen not as a causal factor, but rather as a mediating (contextual) factor in intimate abuse.

References

Bartholomew, K. (1990). Avoidance of intimacy: An attachment perspective. Journal of Social and Personal Relationships, 7, 147-178.

Bartholomew, K., & Horowitz, L. M. (1991). Attachment styles among young adults: A test of a four-category model. Journal of Personality and Social Psychology, 61, 226-244.

Bell, A. P., & Weinberg, M. S. (1978). Homosexualities: A study of diversity among men and women. New York: Simon & Schuster.

Bell, A. P., Weinberg, M. S., & Hammersmith, S. K. (1981). Sexual preference: Its development in men and women. New York: Simon and Schuster.

Berk, R. A., Berk, S. F., Loseke, D. R., & Rauma, D. (1981). Mutual combat and other family violence myths. In D. Finkelhor, R. J. Gelles, G. T. Hotaling, & M. A. Straus (Eds.), The dark side of families: Current family violence research (pp. 197-212). Beverly Hills, CA: Sage.

Blood, R., Jr. & Wolfe, D. (1960). Husbands and wives: The dynamics of married living. Glencoe, IL: Free Press.

Bograd, M. (1988). Feminist perspectives on wife abuse: An introduction. In M. Bograd & K. Yllo (Eds.), Feminist perspectives on wife abuse (pp.11-26). Beverly Hills: Sage.

Browne, A. (1992). Are women as violent as men? Unpublished manuscript, University of Massachusetts, Worcester, MA.

Coleman, D. H., & Straus, M. A. (1986). Marital power, conflict, and violence. Violence and Victims, 1, 141-157.

Collins, N. L., & Read, S. L. (1990). Adult attachment, working models and relationship quality in dating couples. Journal of Personality and Social Psychology, 61, 644-663.

Dobash, R. E., & Dobash, R. P. (1979). Violence against wives: A case against the patriarchy. New York: Free Press.

Dutton, D. G. (1994a). Behavioral and affective correlates of Borderline Personality Organization in wife assaulters. International Journal of Law and Psychiatry, 17, 265-279.

Dutton, D.G. (1994b). Patriarchy and wife assault: The ecological fallacy. Violence and Victims, 9, 167-182.

Dutton, D. G., Saunders, K., Starzomski, A., & Bartholomew, K. (1994). Intimacy-anger and insecure attachment as precursors of abuse in intimate relationships. Journal of Applied Social Psychology, 24, 1367-1386.

Dutton, D. G., & Starzomski, J. A. (1993). Borderline personality in perpetrators of psychological and physical abuse. Violence and Victims , 8, 327-337.

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

Farley, N. (1996). A survey of factors contributing to gay and lesbian domestic violence. Journal of Gay & Lesbian Social Services, 4, 35-42.

Gelles, R. J., & Straus, M. A. (1988). Intimate violence. New York: Simon and Shuster.

Gerslma, A., Emmelkamp, P. M. G., & Arrindell, W. A. (1990). Anxiety, depression and perception of early parenting: A meta-analysis. Clinical Psychology Review, 10, 251-277.

Gonzalez, R., & Griffin, D. (1996). On the statistics of interdependence: Treating dyadic data with respect. In S. Duck (Ed.), Handbook of personal relationships (2nd ed., pp. 271-302). Chichester, New York: John Wiley and Sons.

Gray-Little, B., & Burk, N. (1983). Power and satisfaction in marriage: A review and critique. Psychological Bulletin, 93, 513-538.

Griffin, D. W., & Bartholomew, K. (1994). The metaphysics of measurement: The case of adult attachment. In K. Bartholomew & D. Perlman (Eds.), Advances in personal relationships: Attachment processes in adulthood (Vol. 5, pp. 17-52). London, England: Jessica Kingsley.

Gunderson, J. G. (1984). Borderline Personality Disorder. Washington, DC: American Psychiatric Press.

Hamberger, L. K., & Hastings, J. E. (1986). Characteristics of male spouse abusers: Is psychopathology part of the picture? Paper presented at American Society of Criminology, Atlanta, GA.

Hamberger, L. K., & Hastings, J. E. (1991). Personality correlates of men who batter and nonviolent men: Some continuities and discontinuities. Journal of Family Violence, 6, 131-147.

Hamberger, L. K., & Potente, T. (1994). Counseling heterosexual women arrested for domestic violence: Implication for theory and practice. Violence and Violence, 9, 125-137.

Harry, J. (1982). Gay children grown up: Gender culture and gender deviance. New York: Praeger.

Harry, J. (1984). Gay couples. New York: Praeger Publishers.

Harry, J. (1989). Parental physical abuse and sexual orientation in males. Archives of Sexual Behavior, 18, 251-261.

Harry, J. (1990). A probability sample of gay males. Journal of Homosexuality, 19, 89- 104.

Hart, S. D., Dutton, D. G., & Newlove, T. (1993). The prevalence of personality disorder among wife assaulters. Journal of Personality Disorders, 7, 328-340.

` Hazan, C., & Shaver, P. (1987). Conceptualizing romantic love as an attachment process. Journal of Personality and Social Psychology, 52, 511-524.

Herman, J. L., Perry, J. C. & van der Kolk, B. A. (1989). Childhood trauma in Borderline Personality Disorder. American Journal of Psychiatry, 146, 490-495.

Island, D., & Letellier, P. (1991). Men who beat the men who love them. New York: Harrington Park Press.

Kasian, M., & Painter, S. (1992). Frequency and severity of psychological abuse in a dating population. Journal of Interpersonal Violence, 7, 350-364.

Kelly, E. E., & Warshafsky, L. (July, 1987). Partner abuse in gay male and lesbian couples. Paper presented at the Third National Conference for Family Violence Researchers, Durham, NH.

Kenny, D. A. (1988). The analysis of data from two-person relationships. In S. Duck (Ed.), Handbook of personal relationships: theory, research, and interventions (pp. 57-77). New York: John Wiley and Sons.

Kenny, D. A. (1996). Models of non-independence in dyadic research. Journal of Social and Personal Relationships, 13, 279-294.

Keppel, G. (1991). Design and analysis: A researcher’s handbook. Englewood Cliffs, New Jersey: Prentice Hall.

Kernberg, O. (1977). The structural diagnosis of borderline personality organization. In P. Hartocollis (Ed.), Borderline Personality Disorder: The concept, the syndrome, the patient (pp. 87-121). New York: International Universities Press.

Letellier, P. (1994). Gay and bisexual male domestic violence victimization: Challenges to feminist theory. Violence & Victims, 9, 125-136.

Martin, D. (1976). Battered wives. New York: Simon & Schuster.

McWhirter, D. P., & Mattison, A. M. (1984). The male couple. Englewood Cliffs, NJ: Prentice-Hall.

Merrill, G. S. (1996). Ruling the exceptions: Same-sex battering and domestic violence theory. Journal of Gay & Lesbian Social Services, 4, 9-21.

Oldham, J., Clarkin, J., Appelbaum, A., Carr, A., Kernberg, P., Lotterman, A., & Haas, G. (1985). A self-report instrument for Borderline Personality Organization. In T. H. McGlashan (Ed.), The Borderline: Current empirical research. The progress in psychiatry series (pp. 1-18). Washington, DC: American Psychiatric Press.

Olson, D. H., & Cromwell, R. E. (1975). Power in families. In R. E. Cromwell, & D. H. Olson (Eds.), Power in families (pp. 15-37). New York: Sage Publications.

Peplau, L. A. (1993). Lesbian and gay relationships. In L. D. Garnets & D. C. Kimmel (Eds.), Psychological perspectives on lesbian and gay male experiences (pp. 395-419). New York: Columbia University Press.

Peplau, L. A., & Gordon, S. L. (1983). The intimate relationships of lesbian and gay men. In E. R. Allgeier & N. B. McCormick (Eds.), The changing boundaries: Gender roles and sexual behavior (pp. 226-244). Palo Alto, CA: Mayfield.

Perris, C., Jacobsson, L., Lindstrom, H., von Knorring, L., & Perris, H. (1980). Development of a new inventory for assessing memories of parental rearing behaviors. Acta Psychiatrica Scandinavia, 61, 265-274.

Reece, R., & Segrist, A. E. (1981). The association of selected "masculine" sex-role variables with length of relationship in gay male couples. Journal of Homosexuality, 7, 33-47.

Rosenthal, R., & Rosnow, R. L. (1975). The volunteer subject. New York: John Wiley & Sons.

Ross, M. W., Campbell, R. L., & Clayter, J. R. (1982). New inventory of measurement of parental rearing patterns: An English form of the EMBU. Acta Psychiatrica Scandinavica, 66, 499-507.

Saghir, M. T., & Robins, E. (1973). Male and female homosexuality: A comprehensive investigation. Baltimore, MD: Williams & Wilkins.

Saunders, D. G. (1988). Other "truths" about domestic violence: A reply to McNeely and Robinson-Simpson. Social Work, 33, 179-183.

Siegel, J. M. (1986). The multidimensional anger inventory. Journal of Personality and Social Psychology, 51, 191-200.

Silverstein, C. (1988). The borderline personality disorder and gay people. Journal of Homosexuality, 15, 185-211.

Smith, M. (1990). Patriarchal ideology and wife beating: A test of the feminist hypothesis. Violence and Victims, 5, 257-273.

Stets, J., & Straus, M. (1990). Gender differences in reporting marital violence and its medical and psychological consequences. In M. Straus & R. Gelles (Eds.), Physical violence in American families. New Brunswick, NJ: Transaction.

Straus, M. A. (1979). Measuring family conflict and violence: The conflict tactics scale. Journal of Marriage and the Family, 41, 75-88.

Symons, D. (1980). The Evolution of human sexuality. Cambridge, MA: Cambridge University Press.

Tolman, R. M. (1989). The development of a measure of psychological maltreatment of women by their male partners. Violence and Victims, 4, 159-177.

van der Kolk, B. (1987). Psychological trauma. Washington, DC: American Psychiatric Press.

Walker, L. E. (1979). The battered woman. New York: Harper & Row.

Walker, L. E. (1984). The battered woman syndrome. New York: Springer Publishing.

Walker, L. E. (1989). Psychology and violence against women. American Psychologist, 44, 695-702.

Whitam, F. L. (1977). Childhood indicators of male homosexuality. Archives of Sexual Behavior, 6 , 89-96.

Yllo, K., & Straus, M. (1990). Patriarchy and violence against wives: The impact of structural and normative factors. In M. Straus & R. Gelles (Eds.), Physical violence in American families (pp.). New Brunswick, NJ: Transaction Publishers.

Zimmerman, M., & Coryell, W. (1989). DSM-III personality disorder diagnosis in a nonpatient sample. Archives of General Psychiatry, 46, 682-689.

Author Note

Monica A. Landolt, Department of Psychology, University of British Columbia; Donald G. Dutton, Department of Psychology, University of British Columbia.

We would like to thank Kim Bartholomew and Dan Perlman for their helpful comments on an earlier draft of this article. We would also like to thank Dale Griffin and Roger Tweed for statistical consultation.

Correspondence concerning this article should be addressed to Monica A. Landolt, Department of Psychology, University of British Columbia, 2136 West Mall, Vancouver, British Columbia, V6T 1Z4, Canada. Electronic mail may be sent via Internet to mlandolt@unixg.ubc.ca.

Table I

Self (Actor)-Reported Partner Reported Abuse

Personality Characteristics Psychological Physical

*p < .05, two-tailed. **p < .01, two-tailed. ***p < .001, two-tailed.

Table II

Psychological Abuse Physical Abuse

*p < .05, two-tailed. **p < .01, two-tailed. ***p < .001, two-tailed.

Figure Caption

Figure 1. Mean scores of actor perpetrated and victim perpetrated abuse as a function of victims’ perception of power.

Figure 2. The centrality of BPO in a sample of gay males.