Abstract
Background:
Nurturing environment contributes to the personality development of young people, but how inappropriate parental bonding from biological parents influences the personality disorder functioning styles of their offspring in divorced families remains unclear.
Objective:
We aimed to find out the relationship between parenting styles and students’ disordered personality in divorced families.
Methods:
The Parental Bonding Instrument (PBI) and the Parker Personality Measure (PERM) were administered to 60 Chinese students from divorced family (Divorced group) and 120 controls (matched in age and gender) from intact family (Intact group). The depressive tendency of the participants was measured with the Plutchik–van Praag Depression Inventory (PVP). Results: PERM borderline style was more prominent when father was perceived as secondary carer than primary carer in Divorced group. When father was perceived as secondary carer, PERM borderline style was more obvious in Divorced group than in Intact group. Maternal autonomy denial and care predicted PERM narcissistic and dependent styles in Divorced group, when mother was perceived as primary carer. In the Intact group, when mother was perceived as secondary carer, maternal care was lower than when mother was perceived as primary carer. Some parental PBI scales predicted certain PERM scales regardless of the carer order.
Conclusions:
The results imply that in divorced family, biological father should be more involved in young people’s daily life, and biological mother should show more care and less autonomy denial. In intact family, more care, less freedom control, and less autonomy denial from both parents are important in the personality development of their offspring.
Keywords
Parental bonding styles, such as care, freedom control, and autonomy denial, contribute significantly to the development of the psychological world of offspring (Enns, Cox, & Clara, 2002; Gao, Raine, Chan, Venables & Mednick 2010). For instance, lack of maternal care and low paternal overprotection are related to the behavioral and emotional problems of the offspring during their childhood or later adulthood (Silberg & Eaves, 2004), such as anxiety (Rork & Morris, 2009), depression (Narita et al., 2000), and internalizing and externalizing problems (Campbell, Pierce, Moore, Marakovitz, & Newby, 1996; Deković, 1999; Enns et al., 2002; Hartman, Stage, & Webster-Stratton, 2003). Low maternal care and low paternal/maternal overprotection were associated with the aggressive and delinquent behaviors in children and adolescents (Hiramura et al., 2010). In addition, lack of parental care, overprotection, or both in childhood had resulted in the change of normal personality such as higher neuroticism, self-criticism, and perfectionism (Enns, Cox, & Larsen, 2000), and in the development of antisocial or psychopathic personality disorder (Gao et al., 2010; Reti et al., 2002) in adulthood. On the other hand, personality disorder patients perceived less parental care and more paternal freedom control and autonomy denial than did normal adolescents and adults (Yu et al., 2007; Zheng et al., 2011).
Family Environment and Offspring’s Psychopathology
Moreover, environmental factors, such as parental marital conflict, and structures of the intact, divorced, or adopted family also affect the cognitive, social, and emotional well-being of young people worldwide (reviewed in Amato 2005; Cowan & Cowan, 2002). The standard family environment model has been proposed to illustrate that marital conflict and divorce, increasing the risk of young people’s behavior problems (Amato & Cheadle, 2008). For example, experiencing parental relationship instability in early childhood was associated with the sexual behavior and major depression in adolescents (Donahue et al., 2010). The resolution of the interfamilial conflict such as reasoning, verbal aggression, or violence was also associated with depressive symptomatology in adolescents (Kashani, Burbach, & Rosenberg, 1988). Marital conflict concerning children and adolescents predicted their behavioral changes, particularly in stepfamilies (Jenkins, Simpson, Dunn, Rasbash, & O’Connor, 2005). On the other hand, a prospective study has shown that a family adversity, such as hostility, resentment, and hitting/shouting between parents, has a direct impact on the borderline personality disorder symptoms of children, leading to the diagnosis of borderline personality disorder in adulthood (Winsper, Zanarini, & Wolke, 2012). The event and process of parental divorce were important factors contributing to the disordered personality traits in children, adolescents, and young adults, especially those from the divorced family who were more aggressive, impulsive, and antisocial (Kelly, 2000).
When regarding the divorced family, however, exact relationships between different types of personality disorders and different styles of parental bonding are still missing. Children in divorced family might live with their single parents, stepparents, grandparents, or other relatives, where they would perceive special parental bonding styles. In intact family, both biological parents might take care of their offspring, while in divorced family, one parent might pay more attention to the young people than another parent does, or none of them takes care of the offspring. In order to achieve a homogeneous comparison between intact and divorced families, we would only include the perceived parenting styles of the biological parents who were viewed as the first- or the second-intimate carers by our participants.
Carer Order and Parental Bonding
Based on previous reports, we would like to hypothesize that in divorced family, paternal and maternal bonding styles would have different effects on the personality disorder functioning styles of their offspring, while the effects would be different in intact families, and the carer order (primary vs. secondary) would affect the parental bonding styles’ influences on the personality disorder functioning styles. In the current study, we would like to apply the Parker Personality Measure (PERM; Parker & Hadzi-Pavlovic, 2001) to assess 11 personality disorder functioning styles (paranoid, schizoid, schizotypal, antisocial, borderline, histrionic, narcissistic, avoidant, dependent, obsessive-compulsive, and passive-aggressive), and the Parental Bonding Instrument (PBI; Parker, Tupling, & Brown, 1979) to measure care, freedom control, and autonomy denial of the carers perceived by the participants. Moreover, we would like to apply the Plutchik–van Praag depression inventory (PVP; Plutchik & van Praag, 1987) to measure the depressive tendency of these participants.
Method
Participants
The participants were 60 students from divorced families (Divorced group: 32 female and 28 male students, aged 18.3 years with 2.8 SD, range 14–25 years) and 120 from intact families (Intact group: 64 female and 56 male students, aged 17.9 years ± 2.2, range 15–24 years), who were recruited from a psychology class and paid for course credit. The students enrolled here included both adolescent and young adult participants, and they were free from any neurological lesions or any active psychiatric disorder, free from alcohol or drug use at least 72 hr prior to participating in the study. Participants with a PVP score higher than 25 (indicating a depressive state, see below) were also excluded. In addition, they had given their informed written consents to be included in the study. This study has been approved by a local ethics committee.
Measures
Participants were asked to sort the order of their first four carers according to their relationship intimacy with them, from the most intimate one (designated as primary carer) to the least (fourth carer). A carer might be the guardian or custodian in the divorced family but not always. Each participant underwent the following three inventory tests:
PBI (Parker et al., 1979). This is a self-report assessing perceived parenting before age 16 with 25 items. The test provides measures of three factors, namely, care, freedom control, and autonomy denial (Table 1). Each PBI item consists of a 4-point Likert-type scale (1 = very unlike me, 2 = moderately unlike me, 3 = moderately like me, and 4 = very like me). The Chinese version of PBI was tested to be reliable in a previous study (internal αs: .70–.88; Yu et al., 2007). Participants completed four sets of PBI according to the perceived parenting from their four carers, respectively. The internal reliability of each scale in the present study was listed in Table 1. PERM (Parker & Hadzi-Pavlovic, 2001). PERM measures functioning styles of 11 personality disorders (Table 1). Each item consists of a 5-point Likert-type scale (1 = very unlike me, 2 = moderately unlike me, 3 = somewhat unlike like me, 4 = moderately like me, and 5 = very like me). Higher score indicates an abnormal personality disorder functioning style. The Chinese version of PERM has previously been shown to be reliable in two student samples in China (internal αs: .35–.78; Wang et al., 2003). The internal reliability of each scale in the present study was listed in Table 1. Similar to those in Wang et al. (2003), the internal αs of Schizoid and Obsessive-Compulsive Scales were also lower in the current study, which might be due to the cultural differences regarding to the diagnostic descriptors of those personality disorders (Wang et al., 2003). The PVP (Plutchik & van Praag, 1987). Each of the 34 PVP items is scored 0, 1, or 2, corresponding to an increasing depressive tendency. Participants have “possible depression” if they score between 20 and 25 or “depression” if they score above 25. According to a recent study (Wang, Cao, Zhu, Gu, & Wang, 2002), the internal reliability of the inventory was .94 in a Chinese sample.
Item Numbers and Internal αs of the Plutchik–van Praag Depression Inventory, Parental Bonding Instrument, and Personality Measure in 180 Participant.
Statistical Analyses
Only PBI data regarding biological parents as either primary carer or secondary carer were retained for further analyses (see Results section). Mean scores of each PBI or PERM scale in four subgroups, that is, family (divorced vs. intact) × carer (primary vs. secondary), either father or mother were analyzed by one-way analysis of variance (ANOVA). Once a significant main effect was detected, a post hoc test with the least significant difference was used to evaluate the individual scale score differences. Moreover, in each of the eight subgroups, that is, family (divorced vs. intact) × carer (primary vs. secondary) × parent (father vs. mother), we used the multiple regression analyses between PBI scales (as predictors) and 11 PERM scales through a stepwise procedure. A p value less than .05 was considered as significant.
Results
Most participants perceived their biological parents as their primary and secondary carers. However, in the divorced families, the numbers of participants who perceived grandfather, grandmother, and others as primary carer were 6, 11, and 2, respectively; while those who perceived grandfather, grandmother, and others as secondary carer were 10, 15, and 7, respectively. In intact families, the numbers of participants who perceived grandfather and grandmother as primary carer were 9 and 11, respectively, while those who perceived grandfather and grandmother as secondary carer were 4 and 15, respectively. Due to the small sample sizes of these participants, we excluded the data from them but only retained the data regarding the biological parents as either primary carer or secondary carer for further analyses. All the remaining participants scored less than 25 on PVP. Furthermore, there was no statistically significant difference between subgroups, F(3, 132) = 0.73, mean squared error = 28.36, p = .536, when mother was regarded as carers; F(3, 130) = 2.39, MSE = 107.70, p = .072, when father was regarded as carers, regarding PVP (Table 2).
Scale Scores (Mean ± SD) of the Plutchik–van Praag Depression Inventory, Parental Bonding Instrument, and Personality Measure in Different Subgroups of Participant.
ap < .05 versus mother as primary carer in Intact group. bp < .05 versus father as primary carer in Divorced group. cp < .05 versus father as secondary carer in Intact group.
PBI Scale Scores
Regarding the paternal PBI scale scores in Divorced and Intact groups, one-way ANOVA did not detect any significant differences when either father was treated as primary carer or as secondary carer (Table 2). By contrast, when mother was regarded as carers, mean scores of maternal care were significantly different among subgroups, F(3, 132) = 2.72, MSE = 104.14, p = .047. In Intact group, the mean scores of maternal care when mother was treated as primary carer (27.86 ± 5.65) were significantly higher (95% CI [0.49, 5.44], p = .019) than those when mother was treated as secondary carer (24.89 ± 6.82; Table 2).
PERM Scale Scores
When regarding mean scores of each PERM scale and when father was treated as primary carer or secondary carer, there was one statistically significant difference between subgroups relating to the borderline personality disorder functioning style, F(3, 130) = 2.73, MSE = 135.25, p = .047. The post hoc least significant difference demonstrated that participants who treated father as secondary carer in Divorced group (45.06 ± 6.84) scored significantly higher on the borderline style than those who treated father as primary carer in the same group (39.06 ± 6.02; 95% CI [1.04, 10.97]; p = .018), and higher than those who treated father as secondary carer in Intact group (39.79 ± 7.25; 95% CI [0.88, 9.67]; p = .019). Regarding PERM scale scores when mother was treated as primary carer or secondary carer, however, there were no statistically significant differences between subgroups (Table 2).
Relationship Between PBI and PERM Scales
In Divorced group, the relationships between PBI and PERM scale scores were quite weak, except that when mother was perceived as primary carer, maternal autonomy denial, β = 0.439, t(19) = 2.13, p = .047, predicted the narcissistic personality disorder functioning style, R2 = .19, F(1, 19) = 4.53, p = .047, and both maternal autonomy denial, β = 0.717, t(18) = 3.69, p = .002, and maternal care, negatively, β = −0.563, t(18) = −2.90, p = .009, predicted the dependent style, R2 = .46, F(2, 18) = 7.76, p = .004 (Table 3).
Stepwise Multiple Regression Predicting Personality Disorder Functioning Styles With Parenting Styles in Different Subgroups of Participants.
Note. Only significant regression coefficients (β) with predictors are presented here for clarity. A = autonomy denial; F = freedom control; C = care; R2-a, adjusted R2.
In Intact group, when father was perceived as primary carer, paternal care (negatively) and autonomy denial predicted the schizoid, antisocial, and passive-aggressive styles (R2s from .09 to .10; Table 3). When father was perceived as secondary carer, paternal freedom control and autonomy denial predicted the paranoid, borderline, histrionic, narcissistic, dependent, and passive-aggressive styles (R2s from .10 to .28). When mother was perceived as primary carer, maternal freedom control positively predicted all personality disorder functioning styles except the paranoid (R2s from .10 to .34). When mother was perceived as secondary carer, maternal care negatively predicted the schizotypal, borderline, and dependent styles (R2s from .07 to .11; Table 3).
Discussion
PERM borderline style was more prominent when father was perceived as secondary carer than primary carer in Divorced group. When father was perceived as secondary carer, PERM borderline style was more obvious in Divorced group than in Intact group. Only maternal autonomy denial and care predicted PERM narcissistic and dependent styles in Divorced group when mother was perceived as primary carer. In our Intact group, however, when mother was perceived as secondary carer, maternal care was lower than when mother was perceived as primary carer; some parental PBI scales predicted certain PERM scales regardless of the carer order. Our results do not fully support the conclusion that parental divorce was not a risk factor for young people’s personality disorders (Kantojärvi et al., 2008) but are in line with other findings that their personality traits are indeed influenced to some extent (e.g., Brennan & Shaver, 1998; Distel et al., 2011).
In our Divorced group, although PBI scale scores regarding the biological parents were similar to each other when they were perceived as either primary carer or secondary carer, higher PERM borderline style score was found in the subgroup when father was perceived as secondary carer, suggesting a crucial role for the paternal bonding in the divorced family. The results confirmed the gradual importance of the moderator effects on the relationship between young people’s psychopathology and the paternal involvement in the standard family environment model (Flouri, 2010). Indeed, some nonresident fathers do not have regular contact with their offspring and thus engage in less authoritative parenting, while the latter often promotes their offspring’s welfare (Amato, Sobolewski, & Lamb, 2004). Further, being consistent with the standard family environment model (Flour, 2010), Rohner and Veneziano (2001) have found that whether in divorced or in intact family, the influence of paternal love (care) on the development of the young people is as great as and occasionally greater than the influence of maternal love. Rohner and Brothers (1999) had, in addition, noted that women with borderline personality disorder perceived more paternal rejection.
Dysfunctional early interactions with parents have been viewed as the key to the emergence of narcissism. One main theory holds that parental coldness, extremely high expectations and low support, may lead to narcissism (Horton, Bleau, & Drwecki, 2006; Kernberg, 1975; Kohut, 1977; Stinson et al., 2008; reviewed in Thomaes, Bushman, Orobio de Castro, & Stegge, 2009). Consistently, when mother was perceived as primary carer, maternal autonomy denial predicted the narcissistic style in our Divorced group. Moreover, when mother was perceived as primary carer, maternal autonomy denial and care (negatively) predicted the dependent style in our Divorced group. Generally, low paternal/maternal affection or nurturing (similar to care) was significantly associated with personality disorders including the dependent type (Johnson, Cohen, Chen, Kasen, & Brook, 2006). Lower affectionless control (low care/high autonomy denial) from mothers is associated with the specific personality traits such as the higher harm avoidance or lower self-directedness (similar to dependent), especially in the opposite sex offspring (Otani, Suzuki, Oshino, Ishii, & Matsumoto, 2009; Oshino, Suzuki, Ishii, & Otani, 2007).
Consistent with the perceived carer ordering, participants who regarded mothers as primary carers in our Intact group perceived more maternal care. In Intact subgroup where father was perceived as primary carer, low paternal care predicted schizoid, which confirms the relationship between attachment patterns and schizoid personality disorder (West, Rose, & Sheldon-Keller, 1994) and accords with that paternal rejection (low care) was related to paranoid, schizoid, and schizotypal personality pathologies (Thimm, 2010). In this subgroup, we also found that paternal autonomy denial predicted antisocial and passive-aggressive styles, which was in line with the relationship between paternal overprotection and the delinquent or antisocial behavior in adolescents (e.g., Hiramura et al., 2010; Mak, 1994; Vera, Granero, & Ezpeleta, 2012). In Intact subgroup where father was perceived as secondary carer, paternal autonomy denial predicted the paranoid, borderline, narcissistic, and passive-aggressive styles, and paternal freedom control predicted the histrionic, narcissistic, and passive-aggressive styles. These results were partly consistent with our previous finding that the personality disorder patients perceived higher paternal autonomy denial and freedom control than the healthy volunteers did (Zheng et al., 2011). In Intact subgroup where mother was perceived as primary carer, the maternal freedom control predicted almost all PERM scales except the paranoid style. Indeed, maternal overprotection was associated with many personality disorders, such as the antisocial (Reti et al., 2002), dependent (Russ, Heim, & Westen, 2003), and obsessive-compulsive types (Wilcox et al., 2008). In Intact subgroup where mother was perceived as secondary carer, low care predicted the schizotypal, borderline, and dependent styles, which generally accords with the finding that maternal care was an important predictor of offspring’s personality (e.g., Gao et al., 2010; Russ et al., 2003; Yu et al., 2007). Taken together, our findings support considering both warmth and control as key parenting dimensions (Westbrook & Harden, 2010), which might offer some clues for different psychotherapies applied in different situations.
However, one should also bear in mind some limitations of our study design. Firstly, when considering factors influencing personality, we only focused on parenting patterns of biological parents, not including other social factors or the nurturing patterns of grandparents, stepparents, or other relatives. Secondly, in Divorced group, we did not record the age of participants when their biological parents divorced. Thirdly, the parental bonding scale scores depend on self-report, which are subject to recall bias. Finally, sample sizes for two subgroups were small (less than 20), the statistical power in this regard was weak. Nevertheless, our results imply that in divorced family, although other social factors might influence the development of adolescents’ personality, biological father should be more involved in students’ daily life, and biological mother should show more care and less autonomy denial to them. In addition, in intact family, more care, less freedom control, and less autonomy denial from both parents are called for a healthy personality development of the offspring.
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.
