Traditionally, and somewhat expectedly, the narcissism literature has predominantly focused on men, whereas the study of narcissism in women has received comparatively little systematic investigation (Green et al., 2022; Grijalva et al., 2014). Since its inception, narcissism has been closely associated with the traditional gender socialisation of men due to its core traits of grandiosity as they appear in current clinical diagnostic descriptions (American Psychiatric Association, 2013) and ‘gold standard’ assessments in the social/personality literature (e.g., Narcissistic Personality Inventory; NPI; Raskin & Terry, 1988). Such prototypical traits include preoccupation with omnipotence, exhibition of self-aggrandizing tendencies, pursuit of power, a sense of entitlement, lack of empathy, interpersonal exploitation, and an authoritarian character style; features which are more strongly present in men than women (Weidmann et al., 2023). In the past decade, there has been a surge in empirical literature that recognises the heterogeneity of narcissism, encompassing vulnerable manifestations in addition to grandiosity. In stark contrast to their grandiose counterparts, vulnerable narcissistic individuals display features of inhibition, inadequacy, hypervigilance, shame, low self-esteem, and incompetence (Miller et al., 2011). Despite differences in overt features, the two variations of narcissism differ in negative emotionality and agency, but share an antagonistic core (Miller et al., 2021; Pincus, 2023).

Until recently, theoretical understandings of narcissism in women have been impeded by researchers applying male-coded criteria to women whilst ignoring pronounced gender disparities in the phenotypic expression of this phenomenon (for a review, see Green et al., 2022). This pattern occurs despite research demonstrating the vulnerable type of narcissism to be either gender neutral (Grijalva et al., 2014; Weidmann et al., 2023) or being more prevalent in women (Green et al., 2022; Pincus et al., 2009; Wright et al., 2010). Recent research by Green et al. (2020a) asserted that these gender differences may partially stem from gender-specific trait stereotypes of masculinity and femininity that have been ingrained from early parent-child interactions. In this context, the tendency for men to display more grandiose features and women to display more vulnerable traits may originate from differences in parenting approaches designed to make boys more agentic and girls more communal (Green et al., 2020a). In line with social role theory (Wood & Eagly, 2012), agentic behaviours are more socially adaptive features associated with men, whereas women are likely faced with tougher sanctions when displaying stereotypical features of narcissism. Gendered socialisation practices associated with femininity and masculinity may therefore shape the expression, and preponderance, of vulnerable narcissism in women and grandiose narcissism in men (Green et al., 2020a).

Theoretical perspectives contend that narcissistic features emerge because of early dysfunctional parent-child interactions (see Horton, 2011, for a review). Such maladaptive parental practices may include emotional neglect, physical abuse, overprotectiveness, and overindulgent parenting; all of which have been found to contribute to narcissistic disturbances in the child. Indeed, recent meta-analytic findings add to a growing empirical base of studies demonstrating that exposure to childhood maltreatment is a precursor for the development of grandiose and vulnerable narcissism (see Gao et al., 2024). In turn, research shows that narcissism is a critical risk factor for violence in adulthood (Ménard et al., 2021). For instance, narcissism has been associated with impaired quality of relationships and partner violence perpetration due to the tendency to aggress interpersonally in response to ego-threats, exhibit elevated entitlement to special treatment, lack of empathy, and interpersonal exploitation to achieve own ends (Day et al., 2022b).

Intimate Partner Violence (IPV) – broadly defined as the perpetration of physical, sexual, and psychological abuse towards an intimate partner – is a serious societal concern that poses significant mental and physical health risks for affected individuals (Day et al., 2019, 2022a; Green & Charles, 2019; Green et al., 2019). Despite concerning figures which estimate 2.9 million men have been a victim of domestic violence since the age of 16 (Home Office, 2022), surprisingly little is known about the antecedents of violence perpetrated by narcissistic women. While intimate partner violence is perpetrated predominately by men (Home Office, 2022), without considering the contributing role of narcissism (and particularly vulnerable narcissism), instances of perpetration by women may be missed as it does not correspond with the archetype of a ‘typical’ abuser. Therefore, there is a need to expand our understanding of risk markers of IPV to include those of women with salient features of vulnerable narcissism, in order to improve identification, prevention and intervention efforts.

The present study aims to enhance theory and inform practice regarding the extent to which grandiose and vulnerable narcissism mediate the association between childhood exposure to maltreatment and subsequent partner violence in adulthood, and the moderating role of gender in these associations.

Gender Differences in the Development of Narcissism

The emergence of narcissistic features in adulthood have been largely attributed to environmental factors. Early theorists have ascribed adverse parenting to the development of narcissistic personality traits, specifically referring to neglectful (Kohut, 1977), combined with overprotective (i.e., intrusiveness and controllingness; Kernberg, 1975), and overindulgent (Millon, 1981) parenting. Although empirical examinations into narcissism and childhood experiences have been widely studied, the literature has yielded inconclusive results due to assessments of narcissism (NPI) as a unidimensional concept and singular or multiple assessments of different parenting styles being utilised (for a review, see Kılıçkaya et al., 2023). More recently, research has provided a clearer picture regarding the aetiology of narcissism by considering the gendered pathways in the emergence of grandiose and vulnerable features. For example, Mechanic and Barry (2015) found that recalled perceptions of inconsistent parental discipline predicted unique variance in vulnerable narcissism, with a main effect also present for gender (i.e., women scoring higher). Similarly, Ensink et al. (2017) demonstrated that recalled parental neglect and psychological abuse was positively associated with vulnerable and grandiose narcissism in girls.

Taking into consideration parent gender, Cramer (2015) found that a mother’s parenting practice was associated with vulnerable narcissism and a father’s parenting practice was related to grandiose narcissism. Recollections of parental permissiveness and responsiveness by both mothers and fathers were negatively associated with grandiose and vulnerable narcissism, whereas authoritarian parenting by both mothers and fathers were positively associated with both forms of narcissism. Although this study consisted of a small sample (n = 85), other studies also confer these findings. For instance, Huxley and Bizumic (2017) reported that remembered maternal invalidation positively predicted vulnerable narcissism for participants who experienced low to medium levels of paternal invalidation, while higher levels of paternal invalidation positively predicted grandiose narcissism.

Recent work by Green et al. (2020a) provides further empirical support for the parenting styles theoretically associated with the aetiology of narcissism across gender, through an exploration of recalled neglectful (Kohut, 1977), overprotective (Kernberg, 1975), and indulgent parenting (Millon, 1981) by mothers and fathers. Findings revealed that recalled paternal overprotectiveness positively predicted grandiose and vulnerable narcissism in men, whereas recalled accounts of maternal warmth negatively predicted vulnerable narcissism in women in adulthood. On the one hand, these etiological disparities may indicate differences in the conceptual core of narcissism as proposed by early theorists. For instance, Kernberg (1975) conceived narcissism as centred on grandiosity and aggression, whereas Kohut’s (1977) formulation focused on vulnerability and shame. On the other hand, the observed gendered parenting by mothers and fathers may implicate internalisation of stereotyped behaviour which shapes vulnerable narcissistic features in women and grandiose narcissistic features in men. Overall, emergent research in the literature denotes that the onset of narcissism has different developmental antecedents in men and women.

Gender Differences in Intimate Partner Violence Perpetration

The association between narcissism and violence is well-documented in both clinical and community samples (Bogaerts et al., 2021; Bushman, 2017; Day et al., 2022b; Fatfouta et al., 2015; Green & Charles, 2019; Hepper et al., 2014; Johnson et al., 2000; Kalemi et al., 2019; Krusemark et al., 2018). Theoretically, the link between narcissism and violence has been commonly explained in terms of ‘threatened egotism’ in the social/personality literature (Baumeister et al., 2000) and ‘narcissistic injury’ in the clinical field (Freud, 1914/1957; Kohut, 1977; Logan, 2009). According to both models, narcissistic individuals retaliate with rageful and aggressive behaviours in response to perceived threats towards their self-image, as an attempt to (dys)regulate intolerable emotions, such as shame and humiliation. Empirical support for these theories has, however, been largely based on male samples (Barry et al., 2015; Baumeister et al., 2000; Bogaerts et al., 2021; Bushman, 2017; Bushman et al., 2003; Krusemark et al., 2018; Lobbestael et al., 2014; Mouilso & Calhoun, 2016; Palermo, 2008; Velotti et al., 2020). The literature consequently narrates the profile of a narcissistic offender which closely mirrors a hostile masculine personality, where the aggressor reacts with hostility towards perceived insults to his profound sense of self-worth and superiority, which has wounded his pride and masculinity (Baumeister et al., 2000; Mouliso & Calhoun, 2016).

A similar trend has been indicated in the IPV literature, where women are excluded entirely due to the long prevailing belief that men are more narcissistic and aggressive than women (e.g., Buck et al., 2014; Meier, 2004; Rinker, 2009; Talbot et al., 2015). Whilst narcissism and IPV perpetration have been positively linked to both genders, the few studies which include women ignore gender differences in the personality construct. For example, Blinkhorn et al. (2015, 2016, 2018) claim to extend the literature on narcissism in female offenders whilst utilising grandiose narcissism (NPI) as their main assessment, thereby invoking the belief or implicitly assuming that the male template can be transferable to women (see also Caiozzo at al., 2016; Gormley & Lopez, 2010; Hughes et al., 2020 Lamkin et al., 2017; March et al., 2020; Sharma, 2021). These studies have not considered a gender-equivalent assessment that includes vulnerable features of narcissism, despite frequent indications in the literature suggesting that the outward expression of narcissism differs in women (Campbell & Miller, 2011; Grijalva et al., 2014; Morf & Rhodewalt, 2001; Philipson, 1985; Pincus et al., 2009; Richman & Flaherty, 1988; Wright et al., 2010).

The sparse literature that employs gender-appropriate assessments indicate that female aggression and violence are expressed in more subtle and coercive forms, compared to male violence that is more overt and grandiose in nature. Dyadic research conducted by Ryan et al. (2008) found that, in women only, the exploitative/entitlement facet of grandiose narcissism, which significantly correlated with vulnerable narcissism, was related to sexual coercion. These findings were interpreted as narcissistic women being more hypersensitive to the perceived coercive behaviours of their partners, consequently exerting manipulative strategies to gain control in their relationship. Similarly, Southard (2010) found that, compared to men, vulnerable narcissism and the exploitative/entitlement facet were only related to women’s use of specific manipulative tactics such as bullying and disengagement. Related research finds that both grandiose and vulnerable features in both genders were significantly linked to their perpetration of psychological abuse in intimate relationships (Ponti et al., 2020), whereas other research finds vulnerable narcissism as a significant predictor of cyber intimate partner violence in women (Branson & March, 2021; March et al., 2020b), but not in men (March et al., 2020). Each of these studies are, however, limited in exploring only certain tactics of IPV as opposed to the full spectrum (physical/sexual, and psychological abuse).

A novel qualitative study by Green et al. (2019) enhanced theoretical knowledge regarding narcissism in women and attempts at self-regulation within the full spectrum of IPV. Results showed that gender-related norms shaped motives for women to self-regulate in ego-threatening situations. These strategies were perceived to be obtained through exploiting their feminine qualities (e.g., playing the ‘mother card’, adopting the ‘victim status’) and using legal and societal benefits to assert their dominance over their partner. Essentially, their ‘mask of femininity’ was perceived to resemble features of vulnerable narcissism by intimate partners. Extending these findings in a more comprehensive quantitative study, Green et al. (2020b) found that vulnerable narcissism, but not grandiose narcissism, was the only significant predictor of women’s perpetration of physical/sexual and psychological abuse on a partner. In men, grandiose narcissism predicted psychological abuse and vulnerable narcissism predicted physical/sexual abuse perpetration. These findings support previous speculations that women use more strategic attempts to achieve their narcissistic goals than men, which are not recognised as ‘stereotypically’ narcissistic (Campbell & Miller, 2011; Morf & Rhodewalt, 2001). Thus, the observed gender-specific motives for perpetrating violence in narcissistic individuals may reflect gendered pathways to violence.

Childhood Maltreatment and Subsequent Violence: The Mediating Role of Narcissism

Research shows that those exposed to abuse and adverse parenting practices during childhood are at increased risk of committing violence in adulthood (Ménard et al., 2021). However, there are factors which may mediate this link, given that not all children who are victimised subsequently become violent themselves. Previous research has shown that personality disorders can mediate associations between exposure to childhood maltreatment and later violence, however the literature falls short of sufficiently evaluating the role of narcissism (Brennan, 2014). For instance, Kalemi et al. (2019) reported that narcissism (using the NPI only) and a history of child abuse significantly predicted aggression in a sample of female inmates; however, the study failed to assess the mediating role of narcissism along with the inclusion of vulnerable narcissism. Similarly, Plouffe et al. (2022) explored only grandiose narcissism (as part of the ‘dark triad’ constellation), physical partner violence and a history of early adverse experiences, however, failed to conduct mediation analyses as the relationship between childhood maltreatment and later physical partner violence was non-significant. Including both grandiose and vulnerable narcissism in the evaluation of childhood abuse and subsequent physical partner violence, Ménard et al. (2021) found that neither narcissistic component was found to mediate this relationship. Lastly, Brennan (2014) employed the Pathological Narcissism Inventory (PNI) and found that narcissism partially mediated the link between exposure to child abuse and subsequent general violence in adulthood. Important limitations to note here concerns the operationalisation of narcissism and violence which were treated as unitary assessments, thus gender variations in grandiose and vulnerable narcissism were not explored along with different tactics of violence.

Current Study

As the preceding review of the literature has demonstrated, there is only a nascent literature on narcissism in women using gender-inclusive assessments. This research suggests that the antecedents of narcissism and motives for perpetrating partner violence are differently expressed in women and men, inviting the assumption that there are gendered pathways to offending behaviour that require appropriate interventions. This novel study aims to investigate the mediating role of grandiose and vulnerable narcissism in the association between exposure to childhood adversity and subsequent partner violence, taking into consideration gender differences in these patterns (see theoretical model shown in Fig. 1). Based on previous research, we propose the following hypotheses:

Fig. 1
figure 1

Hypothesised model

Full size image
  1. 1.

    We expect gender differences in narcissism where men score significantly higher in grandiose narcissism than women, and women score significantly higher on vulnerable narcissism than men. This hypothesis is based on the observed longstanding gender differences in narcissism found in past research (e.g., Green et al., 2022; Grijalva et al., 2014).

  2. 2.

    We expect conditional direct effects of retrospective recall of parenting style on self-reported narcissism and IPV perpetration. Specifically, for women, we predict neglectful parenting (lack of care) by the mother will be positively associated with vulnerable narcissism (2a) and later partner violence (2b), for vulnerable narcissism to be positively associated with IPV (2c), and for vulnerable narcissism to mediate the relationship between parenting and IPV (2d; conditional indirect effect). This prediction is based on Kohut’s (1977) theorised parenting style and prior research finding an association between lack of maternal care and vulnerable narcissism in women (Green et al., 2020a), and the link between vulnerable narcissism and IPV in women (e.g., Branson & March, 2021; Green et al., 2019, 2020b; March et al., 2020; Ponti et al., 2020).

  3. 3.

    We also predict that for men, overprotective parenting by the father will be positively associated with grandiose narcissism (3a) and later partner violence (3b), for grandiose narcissism to be positively associated with IPV (3c), and for grandiose narcissism to mediate the relationship between parenting and IPV (3d; conditional indirect effect). This hypothesis is based on Kernberg’s (1975) theorised parenting style and previous research demonstrating an association between overprotective parenting and grandiose narcissism in men (Green et al., 2020a), and the link between grandiose narcissism and IPV in men (e.g., Green et al., 2020b; Meier, 2004; Mouilso & Calhoun, 2016; Rinker, 2009).

Method

Participants

Power analysis software (G*Power 3.1.9.2; Faul et al., 2007) was used to calculate minimum sample size to achieve a desired moderate effect size (f2 = 0.15) at p < .05 significance level using a multiple regression with 19 predictor variables with 80% power: recalled care and overprotective parenting by mother and father (4), participant gender (1), grandiose and vulnerable narcissism (2), and all combinations of parenting*gender (8) and narcissism*gender interactions (4). Power analysis stipulates a minimum of 153 participants is required to achieve a power of 0.80.

From the initial sample pool (n = 704), 371 participants were excluded due to incomplete data. Of those who completed the study (n = 333), five participants were eliminated because they did not identify as any gender (n = 3), were under 18 years old (n = 1), and they did not give informed consent (n = 1). The final analysis was conducted using the remaining 328 participants. The sample comprised 176 (53.7%) women and 152 (46.3%) men. The age range of the participants was 18–64 years with a mean of 27.93 years (SD = 9.09). Relationship status and duration, and stated sexuality, broken down by gender, is displayed in Table 1. The sample was predominantly White (n = 262), with 16 South or East Asian, 12 Hispanic or Latino, 10 African, and five Middle Eastern; the remaining 23 participants chose ‘mixed’ or ‘other’ for their ethnic status.

Table 1 Participant demographics
Full size table

Materials

Pathological Narcissism Inventory

The Pathological Narcissism Inventory (PNI; Pincus et al., 2009) is a 52-item self-report measure of pathological narcissism that assesses both vulnerable (34 items) and grandiose (18 items) features. Responses to the 52-items are made on a 6-point Likert-type scale ranging from 0 (not at all like me) to 5 (very much like me). Seven primary scales of the PNI load on to two higher order domains of Narcissistic Grandiosity and Narcissistic Vulnerability. The scales that load on to Narcissistic Grandiosity include Exploitativeness, Grandiose Fantasy, and Self-Sacrificing Self-Enhancement; and the scales that load on to Narcissistic Vulnerability include Contingent Self-Esteem, Hiding the Self, Devaluing, and Entitlement Rage. Because each subscale varies in scale length, mean item endorsements are used instead of sums to enable ease of comparison across scales (Pincus et al., 2009). The PNI is a widely used measure and manifests good internal consistency (Pincus et al., 2009). In the present study, Cronbach’s alpha for the grandiose component was α = 0.87 and α = 0.95 for the vulnerable component.

Conflict Tactics Scale Short Form

The Conflict Tactics Scale short form (CTS2S; Straus & Douglas, 2004) is a revised 20-item measure of IPV adapted from the longer 39-item measure version of the CTS2. The scale measures perpetration of physical/sexual abuse as well as whether participants have been a victim of violent tactics by their partner. We only focused on the 10-items pertaining to perpetration in this study. The CTS2S uses an 8-point frequency scale to focus on tactics (Negotiation, Psychological Aggression, Sexual Coercion, Physical Assault, and Injury) used during conflict within intimate relationships (0 = this never happened to 8 = this happened more than 20 times in the past). In the current study, in line with Straus and Douglas’ coding scheme, participants were asked to report the occurrence of any violence perpetrated during the course of their relationship, or asked to recall any instances from their most recent relationship. A score of “1” was awarded if one or more acts of violence for each of the tactics during the course of the relationship had occurred and a score of “0” was awarded when no instances of violence were reported. Total ratings were computed for perpetration scores by summing the zeros and ones across the different traits. The CTS2S has demonstrated good construct and concurrent validity (Straus & Douglas, 2004). In the present study, perpetrator reliability was α = 0.69, which is marginally below the proposed cut-off of acceptable internal consistency of 0.70 (Cronbach, 1951). In addition, given that the CTS2S is not designed to sample psychological aggression in depth, it was decided to measure psychological abuse separately.

Multidimensional Measure of Emotional Abuse

The Multidimensional Measure of Emotional Abuse (MMEA; Murphy & Hoover, 1999) is a 28-item scale that specifically measures the psychologically abusive aspect of IPV. Subscales for this questionnaire include Restrictive Engulfment, Denigration, Hostile Withdrawal, and Dominance Intimidation. As with the CTS2S, the MMEA uses an 8-point frequency scale to measure the number of times a particular aspect of emotional abuse has occurred within a relationship. In the current study, participants were asked to report the occurrence of any psychological abuse perpetrated during their relationship or asked to recall any instances from their most recent relationship. The same scoring scale used for CTS2S was adopted for the MMEA inventory to ensure consistent scoring method of prevalence across the IPV questionnaires. The MMEA questionnaire is statistically valid as an index of psychological aggression for research purposes (Murphy & Hoover, 1999). In the present study, internal reliability for perpetration was α = 0.89. Both the MMEA and CTS2S have been utilised in previous research on narcissism and IPV (Carton & Egan, 2017; Green et al., 2020b).

Parenting Bonding Instrument

The Parenting Bonding Instrument (PBI; Parker et al., 1979) measures recollections of parental care (e.g., “He/She was affectionate to me;” “He/She tended to baby me”) and overprotectiveness (e.g., “He/She invaded my privacy;” “He/She tried to control everything I did”). The scale has 12 items reserved for the mother (or female caregiver) and 12 items for the father (or male caregiver). Participants were asked to recall the parenting styles of their parents (or parental figures) during their first 16 years of life on a 4-point rating scale: 1 (very like her/him) to 4 (very unlike her/him). The 12 items for maternal parenting and 12 items for the paternal parenting were totalled to create corresponding indexes. The PBI shows good internal consistency and has been used in previous narcissism research (e.g., Maxwell & Huprich, 2014). In the present study, internal reliability for the total PBI score was α = 0.80.

Procedure

Ethical approval was granted by Edinburgh Napier University School of Applied Sciences Research Integrity Committee. Participants were invited to take part in a study titled “Personality traits and intimate relationships,” which was advertised online on various social media platforms (Facebook, Twitter, Reddit) and research participation websites (psychological research on net), as well as flyers shared at gym facilities which contained a QR code that when scanned, directed participants to the online survey hosted by Qualtrics. Inclusion criteria included being over 18 years of age, being fluent in English, providing informed consent, and experience of being in a relationship. Participants provided informed consent by clicking a box before beginning the survey. They first completed demographic questions and then continued to complete the PNI, CTS2S, MMEA, and the PBI questionnaires, which were presented in that order for each participant. On completion, participants were given the option to enter a draw for a chance to win a £50 Amazon gift voucher. Participants were then directed to the debrief page, thanked, and presented with a list of support networks associated with IPV. Overall, the study took approximately 15–30 min to complete.

Results

Preliminary Analysis

MCAR tests were used to test if missing data were random, and revealed that for the PBI, data were not missing at random; therefore, the mode was used to replace missing data values. Replacing values using the mode is a standard and basic imputation method and, compared to the mean substitution method, does not reduce variance in the dataset (Baraldi & Enders, 2010). All other variables did not show non-random missing data.

Table 2 presents descriptive statistics and Table 3 presents zero-order correlations for the key study variables broken down by gender. Independent samples t-tests (see Table 2) were conducted to test gender differences across all variables for completeness, though we did not have specific predictions for mean gender differences in the study variables, except for the narcissism variables. For parenting style variables, the findings revealed significant gender differences in recall of parental upbringing and exposure to mother care (higher for men), mother overprotectiveness (higher for women), father care (higher for men), and father overprotectiveness (higher for women). For the IPV variables, women reported significantly higher levels of perpetration of physical/sexual abuse and psychological abuse in our sample. There was no significant difference in the length of time the men and women in our sample had been in their current relationship. For our prediction for the narcissism variables, there was no significant gender difference in grandiose narcissism; however, women reported higher levels of vulnerable narcissism compared to men, providing partial support for Hypothesis 1.

Table 2 Independent T-tests on key variables by gender
Full size table
Table 3 Correlations between key variables for men and women
Full size table

Zero-order correlations (Table 3) revealed that for both men and women, there was a negative association between recalled mother and father care and participants’ scores of grandiose and vulnerable narcissism. Mother and father care were also negatively associated with perpetration of physical/sexual and psychological abuse for both men and women. For both men and women, there was a positive association between recalled mother and father overprotectiveness and grandiose and vulnerable narcissism. Mother and father overprotectiveness were positively associated with perpetration of physical/sexual and psychological abuse for both men and women.

As in previous research, grandiose and vulnerable narcissism were positively associated with each other among men and women. Grandiose narcissism was positively associated with psychological, but not physical/sexual, IPV for men, whereas grandiose narcissism was positively associated with both types of IPV for women. Vulnerable narcissism was positively associated with both types of IPV for both men and women.

Finally, relationship duration was negatively associated with vulnerable and grandiose narcissism for women and with grandiose narcissism for men. Relationship duration was thus controlled for in subsequent analyses.

Moderated Mediation

To test Hypotheses 2 and 3, we used PROCESS Version 4.1 to test moderated mediation models (Hayes, 2017; Model 59). We tested whether retrospective reports of mother and father parenting (predictor variables) were associated with grandiose and vulnerable narcissism (mediator variables) and whether retrospective parenting and narcissism were associated with perpetration of two types of IPV: physical/sexual and psychological (outcome variables). We examined all parenting variables simultaneously to allow us to covary out the potential shared variance in exposure to the different parenting styles of both parents. Relationship duration was also included as a covariate. We examined whether all of these relationships were moderated by participant gender (moderating variable).

PROCESS only allows one predictor variable to be entered at a time but can estimate a model with multiple predictor variables by adding these as covariates in the model (Hayes, 2017). Our model had four predictor variables (retrospective reports of mother and father’s care and overprotectiveness). To estimate the direct and indirect effects of the target predictor variable, four models were run for each of the outcome variables. Mathematically, all resulting paths are equivalent to having entered them simultaneously in a structural equation model (Hayes, 2017). All models were run with 10,000 bootstraps. All continuous predictor variables (parenting variables, grandiose and vulnerable narcissism) were centred prior to analysis. Gender was dummy coded such that men = 0 and women = 1. Accordingly, the conditional direct effects reported below use men as the reference group.

Perpetrator Physical/Sexual IPV

Conditional direct effects and interactions are presented in Table 4. Effects broken down by gender and conditional indirect effects are presented in Table 5.

Table 4 Tests of moderated mediation for physical/sexual intimate partner violence
Full size table
Table 5 Conditional effects of focal predictors at values of the moderator (gender) for models examining physical/sexual IPV
Full size table

Retrospective Parenting Style and Gender on Narcissism The betas in Table 4 represent the estimated difference in narcissism scores between two people who differ by one unit in gender amongst those who score at the grand mean on parenting. Across all parenting models (mother and father care and overprotectiveness), gender (with men coded as 0 and women coded as 1) was negatively associated with grandiose narcissism and positively associated with vulnerable narcissism, indicating that men score higher on grandiose narcissism than women and women score higher on vulnerable narcissism than men.

Mother care was negatively associated with grandiose narcissism (when gender = 0) whilst father overprotectiveness was positively associated with participants’ grandiose narcissism.

Narcissism on IPV Across all parenting models, neither grandiose nor vulnerable narcissism was a significant predictor of physical/sexual IPV for men. Table 5 shows that for women, vulnerable narcissism was positively and significantly associated with physical/sexual IPV. For men these effects were also positive but non-significant. Note, however, that the interaction term between vulnerable narcissism and gender across models was non-significant.

Retrospective Parenting Style on IPV Mother care was negatively associated with physical/sexual IPV for men. Mother care was also negatively associated with physical/sexual IPV for women, although this failed to reach statistical significance. The interaction between retrospective reports of mother care and gender was non-significant.

Father overprotectiveness was positively associated with physical/sexual IPV for men. The effect was also positive and significant for women. The interaction term between father overprotectiveness and gender was non-significant.

Indirect Effects of Parenting on IPV via Narcissism Results are presented in Table 5. A significant indirect effect of mother care on physical/sexual IPV via vulnerable narcissism emerged; mother care was negatively associated with vulnerable narcissism and vulnerable narcissism was positively associated with psychological IPV. Mother care thus acted as a buffer against IPV via lower vulnerable narcissism levels. There was no significant indirect effect for men. Note, however that the index of moderated mediation for each model was non-significant, therefore the indirect effects were not related to gender across our models.

We present the above significant results in diagrammatic form in Fig. 2.

Fig. 2
figure 2

Significant gendered paths between parenting, narcissism, and physical/sexual IPV. Note Solid lines represent significant pathways for male participants. Dashed lines represent significant pathways for female participants

Full size image

Perpetrator Psychological IPV

The same model was tested with psychological IPV as the outcome variable. Conditional direct effects and interactions are presented in Table 6. Effects broken down by gender and conditional indirect effects are presented in Table 7.

Table 6 Tests of moderated mediation for psychological intimate partner violence
Full size table
Table 7 Conditional effects of focal predictors at values of the moderator (gender) for models examining psychological IPV
Full size table

Retrospective Parenting Style and Gender on Narcissism The betas in Table 6 represent the estimated difference in narcissism scores between two people who differ by one unit in gender amongst those who score at the grand mean on parenting. Across all parenting models (mother and father care and overprotectiveness), gender (with men coded as 0 and women coded as 1) was negatively associated with grandiose narcissism and positively associated with vulnerable narcissism, indicating that men score higher on grandiose narcissism than women and women score higher on vulnerable narcissism than men. Mother care was negatively associated with grandiose narcissism (when gender = 0) and father overprotectiveness was positively associated with participants’ grandiose narcissism.

Narcissism on IPV Across all models, grandiose narcissism was a positive predictor of psychological IPV (when gender = 0). With the exception of the father care parenting model, significant interactions between vulnerable narcissism and gender also emerged (see Table 7). In all cases, there was a positive association between vulnerable narcissism and IPV for men and women, but these relationships only reached significance for women.

Retrospective Parenting Style on IPV Mother care was negatively associated with psychological IPV for men. Mother care was also negatively associated with psychological IPV for women, although this failed to reach statistical significance. The interaction between retrospective reports of mother care and gender was non-significant.

Father overprotectiveness was positively associated with physical IPV for men. The effect was also positive albeit non-significant for women. The interaction term between father overprotectiveness and gender was non-significant.

Indirect Effects of Parenting on IPV via Narcissism As presented in Table 7, the index of moderated mediation was significant for mother care. That is, for women only there was a significant indirect effect of mother care on psychological IPV via vulnerable narcissism; mother care was negatively associated with vulnerable narcissism and vulnerable narcissism was positively associated with psychological IPV. Mother care thus acted as a buffer against IPV via lower vulnerable narcissism levels. There was no significant indirect effect for men.

The index of moderated mediation values for all other models were non-significant, therefore the indirect effects were not related to gender across these models.

We present the above significant results in diagrammatic form in Fig. 3.

Fig. 3
figure 3

Significant gendered paths between parenting, narcissism, and psychological IPV. Note Solid lines represent significant pathways for men. Dashed lines represent significant pathways for women

Full size image

Discussion

The present study sought to investigate the mediating role of grandiose and vulnerable narcissism in the relationship between recalled parenting practices in childhood and perpetration of partner violence in adulthood, and the extent to which gender moderates these associations. This area is worthy of investigation due to the need for a tailored approach to address risk markers of violence against men by women, which is currently overshadowed by dominant perspectives of male perpetration of violence.

Gender Differences

In line with Hypothesis 1, independent t-tests revealed significant gender differences in mean levels of vulnerable narcissism only (with higher scores for women than men) but no significant difference in levels of grandiose narcissism between men and women. However, all moderated mediation models (which partialled out the effects of the other type of narcissism within the models) revealed gender differences, such that men scored higher on grandiose narcissism than women and women scored higher in vulnerable narcissism than men. These findings are congruent with the vast literature reporting consistent gender disparities in narcissism facets (Green et al., 2020a, b, 2022; Grijalva et al., 2014; Pincus et al., 2009; Weidmann et al., 2023; Wright et al., 2010). As previously theorised, prototypical expressions of narcissism in men and women may be partly symptomatic of prescribed sociocultural norms along masculine and feminine lines (Wood & Eagly, 2012). The tendency for women to align more closely with narcissistic vulnerability and for men to exhibit overt grandiosity may indicate the conformity of such behaviours with cultural gender roles that resemble stereotypical characteristics of women (e.g., low self-esteem, shame, hypersensitivity, neuroticism) and men (e.g., inflated self-image, assertiveness, authority, superiority; Green et al., 2022).

Gendered Roots

Results further revealed significant gender differences when exploring different paths within the models. For the zero-order correlations, recollections of an overprotective parent (mother or father) were positively associated with the development of both forms of narcissism across gender. These results are in concordance with the ideas espoused by Kohut (1977) and Kernberg (1975) who implicated the development of narcissistic features as the result of cold and overprotective parenting, respectively, and in line with past empirical findings (Ensink et al., 2017; Green et al., 2020a; Huxley & Bizumic, 2017). Supporting Hypothesis 2a, retrospective accounts of a caring mother were negatively associated with vulnerable narcissism in women, in line with prior research (Green et al., 2020a), as well as to grandiose narcissism in men. In addition to specific parenting tactics, our findings support the contention that expressions of grandiosity and vulnerability in men and women may be further influenced by the gender of the parent. When controlling for shared variance for all variables, we found that recalled paternal overprotectiveness was associated with higher levels of self-reported grandiose narcissism in men only, consistent with Hypothesis 3a and previous findings (Green et al., 2020a).

Accordingly, mothers and fathers may reinforce stereotyped gendered behaviours associated with agentic behaviours, such as assertiveness and grandiose fantasies in boys, whereas girls are more likely to internalise communal features associated with vulnerability and hypersensitivity (Wood & Eagly, 2012). This might be a reason why paternal overprotectiveness was a significant predictor of grandiose narcissism, but not vulnerable narcissism, in men. These findings may suggest that men have a more complex relationship in their recollections of early life experiences with their fathers, and/or that a father’s role may be more central to their development of (grandiose) narcissism than their mother’s role. Although the self-report methodology precludes substantial confidence in this conclusion, it is nevertheless a possibility that lends itself to further exploration, and, more importantly, underscores the importance of including reports of both parents in future research. This is particularly in light of, and contrary to, the gendered vocabulary articulated (i.e., referring to the parent as mother) when discussing narcissistic development between the child and primary care giver (Freud, 1914/1957; Phillipson, 1982; see Horton, 2011, for an overview).

Gendered Routes

As for associations with partner violence perpetration at the bivariate level, our results showed that both forms of narcissism were positively significantly associated with physical/sexual and psychological perpetration towards an intimate partner in women. In men, only grandiose features were positively significantly associated with psychological perpetration (consistent with Hypothesis 3c) whereas vulnerable features were positively significantly correlated with all forms of IPV perpetration (consistent with Hypothesis 2c). This was unsurprising given past empirical findings linking narcissism facets to IPV (see Green et al., 2022, for a review), and contributes further theoretical support for the ‘narcissistic injury’ premise (Baumeister et al., 2000; Freud, 1914/1957; Kohut, 1977; Logan, 2009). When accounting for shared variance for all other variables, results revealed gender-specific routes to IPV. Here, grandiose narcissism was found to be the only significant predictor of psychological IPV for men, whereas vulnerable narcissism was the only significant predictor of psychological IPV perpetration in women. The need for some men to maintain a grandiose self-image, engage in self-sacrificing self-enhancement attitudes and an exploitative interpersonal style is associated with a greater likelihood of subjecting partners to psychologically abusive tactics. These findings resonate with previous research that found a positive association between grandiose narcissism and the perpetration of psychological abuse (Caiozzo et al., 2016; Carton & Egan, 2017; Gormley & Lopez, 2010; Green et al., 2020b; Peterson & DeHart, 2014; Rinker, 2009).

In contrast, and in line with past empirical literature (Branson & March, 2021; Green et al., 2019, 2020b; March et al., 2020b; Ponti et al., 2020), women’s tendency to express the more covert and feminine-typed traits of narcissism (e.g., hiding the self, fluctuation in self-esteem, rejection sensitivity, devaluation, shame over unmet needs) cultivates a sense of narcissistic entitlement to psychologically abuse intimate partners. As these attributes diverge from the overt masculine-stereotyped traits that comprise grandiose narcissism typically viewed in men, risk markers of narcissism and violence in women may be overlooked given outward expressions of shyness, sensitivity, and insecurity. Thus, the evaluation of narcissism and violence in women is important for theoretical and practical reasons, as self-regulatory motives to obtain power and dominance within interpersonal context appear gender specific. These results arguably significantly limit the generalisability and applicability of research which conceptualises women’s narcissism using male-criteria (e.g., Blinkhorn et al., 2015, 2016, 2018; Caiozzo at al., 2016; Gormley & Lopez, 2010; Hughes et al., 2020; Lamkin et al., 2017; March et al., 2020; Sharma, 2021).

Furthermore, we expected lack of maternal care to significantly predict later partner violence in women (Hypothesis 2b), with vulnerable narcissism mediating this relationship (Hypothesis 2d). Whilst we found no direct effect of maternal care on either type of IPV in women, we did find significant indirect effects: retrospective accounts of a caring mother negatively predicted vulnerable narcissism which was, in turn, indirectly associated with lower instances of psychological abuse towards a partner. It is not surprising that recalled memories of a warm and nurturing mother during childhood fosters an independent self-regard and healthy development of the child’s personality, which in turn leads to more stable adult relationships. We also surmised that lack of paternal overprotectiveness would positively predict IPV in men (Hypothesis 3b), with grandiose narcissism mediating this relationship (Hypothesis 3d). Although we found significant direct effects of paternal overprotectiveness on both forms of IPV, we did not find significant indirect effects, despite the gendered pathways being in the expected directions: for men only, paternal overprotectiveness significantly predicted grandiose narcissism, and grandiose narcissism was positively associated with psychological IPV.

Overall, the current results show gendered roots in the manifestation of narcissism and gendered routes to IPV perpetration. In women only, one novel finding emerged which revealed that recalled maternal care acted as a buffer against IPV via lower vulnerable narcissism levels. These findings stress the need for gender-inclusive interventions to address risk factors in narcissism and IPV perpetration.

Limitations and Future Research Directions

There are several limitations with the current study. For example, data relied on retrospective reports of childhood experiences, thus the possibility that the findings reflect differences in recollection rather than differences in original childhood experience must be acknowledged. However, to do so conclusively would require much more extensive longitudinal research with multiple measures gathered from children’s perspectives of their parent’s parenting practices, along with their parent’s own perspectives on their child-rearing practices. Another issue with retrospective reports in general, and in narcissism research especially, is their propensity to introduce bias in reconstructive memory processes (Morf & Rhodewalt, 2001). Although bias may be present to some extent, childhood recollections provide an important and well-validated first line of evidence into adult consequences of childhood experiences (Chipman et al., 2000). Moreover, potential parent-child interactions could not be directly investigated in the current study, thus the possibility cannot be ruled out that the direction of causality may be either bidirectional or reversed. This is a potential avenue for future research to explore, particularly considering that some research indicates discrepancies exist between parents and adolescents’ views of parenting behaviours assessed (Mechanic & Barry, 2015).

It is also important to note that, whilst clinical theories suggest narcissism emerges as a result of the parent’s narcissistic use of the child, research has found that narcissism is a moderately heritable personality trait and is partly rooted in early emerging temperamental traits (Vernon et al., 2008). Therefore, some children, because of their temperamental traits, might be more likely than others to become narcissistic when exposed to certain environmental stimuli (Miles & Francis, 2014; Thomaes et al., 2009). Future research should study longitudinally the bidirectional association between parenting and adolescent narcissism via genetic influences on parenting as this may account for child characteristics, which could elicit certain parental responses (see Ayoub et al., 2018; also see Klahr & Burt, 2014).

A further limitation pertains to the physical/sexual abuse inventory (CTS2S; Straus & Douglas, 2004) which captures sexual aggression with only two items. It is recommended that future research use a more robust measurement that captures these elements in more depth. Future research should also consider exploring narcissism and IPV in dyadic relationships, and how different sexual orientations and IPV bidirectionality impact gendered expressions of narcissism. It is also worth noting that the items captured by the CTS2S may be skewed towards men and thus do not capture the ways in which women enact physical/sexual IPV. The influence of gendered socialisation and gender norms may have further impacted the (under)reporting and (mis)interpretation of female perpetrated IPV. In other words, women in the current study may be less inclined to admit to overt physical and sexual violence as such acts go against long-grained stereotyped expectations of their feminine gender identity.

Considering the speculations pertaining to gendered parenting in the development of narcissism, future research could also conduct further analysis to examine whether current results are replicated across different family structures (single parent, same-sex parent families) and gender-specific processes. Moreover, research undertaken with parents demonstrates associations between grandiose narcissism and an increased propensity towards non-optimal parenting styles (authoritarian and permissive), with low empathy predicting unresponsive caregiving towards a child (Hart et al., 2017). Given the detrimental ramifications dysfunctional parenting could have on the development of the child, future research could extend these findings to parents with both grandiose and vulnerable narcissism traits, whilst including the role of empathy to assist in the development of effective interventions (see Hart et al., 2017).

Practical Implications

The current findings stress the importance of early parent-child interaction therapy and interventions aimed at reinforcing positive parenting styles and child-rearing environments to ensure healthy attachment in the child and prosocial behaviour. However, notwithstanding the importance of early intervention efforts, the clinical reality is that most patients attend to treatment as adults once these early experiences are entrenched as part of a wider constellation of personality pathology. It is for this reason that contemporary evidence-based treatments for personality disorder pay close attention to the remembered early childhood experiences of adult patients in order to understand an individual’s psychological building blocks – this includes concepts such as the ‘invalidating environment’ of dialectical behaviour therapy (Linehan, 2015), or ‘internalised object relations’ as related to early parent-child interactions for transference focused psychotherapy (Clarkin et al., 2006). These remembered experiences are important for understanding the aetiology of personality pathology, but more importantly also for how such themes are also still active within the patient’s life in the here and now when it comes to areas of intrapersonal and interpersonal difficulty.

Based on the findings of this study, when working with men with elevated narcissistic features there may be a need to attend to current themes of ‘overprotectiveness’ (e.g., intrusiveness, dominance) that they experience from others in their everyday life (and even directly with the therapist). This needs to be explored, including the resulting emotional reactions (e.g., hostility, resentment), potential links with defensive reactions such as superiority and omnipotent control (narcissistic grandiosity) and any concomitant antagonistic and abusive interpersonal patterns towards others. Similarly, when working with women with elevated narcissistic features there may be a need to attend to current themes of being ‘uncared for/unappreciated’ in their everyday life (and even directly with the therapist). This similarly needs to be explored, including any resulting emotional reactions (e.g., shame, envy, entitlement rage), how this relates to a metanarrative of personal victimhood (narcissistic vulnerability) which may then be used to justify antagonistic and abusive interpersonal patterns towards others.

Outlining such internal working models, emotional reactions and subsequent dysfunctional interpersonal patterns are central to effective treatment of personality disorder, including narcissistic personality disorder (Diamond et al., 2021), which then serves as the basis for intervention efforts. As such, given these identified links between gender, narcissistic functioning and intimate partner violence, our findings underscore the importance of assessing and managing risks of interpersonal violence and abuse when working with men and women with prominent narcissistic features as a standard component of clinical care.

Conclusion

In sum, the current findings contribute to the scarce literature on narcissism in women and highlights important gendered roots of the personality construct and gender-specific routes to IPV. Specifically, we found that maternal care negatively predicted vulnerable narcissism in women and grandiose narcissism in men, with paternal overprotectiveness also positively predicting grandiose narcissism in men, but not in women. Grandiose narcissism was a significant predictor of psychological IPV in men, whereas vulnerable narcissism significantly predicted physical/sexual and psychological IPV in women. A novel finding emerged where mother care was associated with reduced psychological IPV via lower vulnerable narcissism levels in women only. The implications of this study raise questions about the vast literature which predominantly portrays men as narcissistic and excludes women. Future research on narcissism should employ gender-inclusive assessments of narcissism that captures vulnerable features to enhance our theoretical understanding of this phenomenon in women. Interventions that target IPV in narcissistic perpetrators can be guided by the gendered risk markers outlined in the current study.