How borderline personality characteristics affect adolescents’ life satisfaction: The role of rejection sensitivity and social relations

Abstract

The present study investigated the association between Borderline Personality characteristics (BP characteristics) and life satisfaction in a sample of 111 girls and 144 boys between the ages of 15 and 19 years. Mediation and moderation effects of rejection sensitivity and adolescents’ social relations with their parents and friends on this association were analyzed. Analyses revealed that more BP characteristics in both boys and girls, and more conflict with parents for girls predicted lower satisfaction with life. Furthermore, conflict with parents mediated and rejection sensitivity moderated the association between BP characteristics and life satisfaction, but again only for girls. Implications of these findings suggest the importance of family conflict management and a gender-specific approach in adolescents with BP characteristics.

Keywords:

• Borderline personality characteristics
• rejection sensitivity
• life-satisfaction
• social relations
• adolescence

Throughout scientific history, people have tried to distinguish between the normal and the abnormal, but have not been able to conclusively make this distinction when considering personality. Gaining a deeper understanding of pathological personality characteristics and their interaction with other individual or contextual factors might help explain why some individuals are more negatively affected by these characteristics than others in their daily lives. This study examined the association between pathological personality characteristics indicative of Borderline Personality Disorder (BPD) and adolescents’ satisfaction with life. Two important factors that have been shown to be related to BPD, and that may lead to important consequences for personal functioning and development, are rejection sensitivity and social relations. These factors are proposed to influence, diminish or exacerbate the effect of Borderline Personality characteristics (BP characteristics) on satisfaction with life, and as such act as mediators or moderators of this relationship.

BP Characteristics and Life Satisfaction

Based on the Diagnostic and Statistical Manual of Mental Disorders (5th ed.; DSM–5; American Psychiatric Association, 2013), BPD can be diagnosed as a pervasive pattern of instability in interpersonal relations, self-image, affects, and behavior. BP characteristics are assumed to be dimensionally distributed throughout the population along a psychopathological severity continuum. Trull and colleagues (2010) consider emotional instability, lack of self-control, and interpersonal hypersensitivity as underlying characteristics of BPD. The DSM-5 proposes a prototype matching system in which personality disorders are defined using narrative descriptions. The borderline personality narrative considers a fragile self-concept or lack of identity, instability in interpersonal relations, and poor emotional regulation as underlying characteristics (Samuel et al., 2012). In line with the dimensional approach to personality pathology, the present study focused on these characteristics in a non-clinical sample of adolescents. BPD usually emerges during adolescence and early detection might be crucial for minimizing possible severe consequences (Chanen, Jovev, & Jackson, 2007). However, little research has been conducted on the underlying mechanisms and consequences of the difficulties related to BP characteristics during adolescence in ‘healthy’ individuals. Some of the consequences of high levels of BP characteristics are that they may interfere with psychological adjustment, attainment of life goals, and personal well-being (e.g., Cramer, Torgersen, & Kringlen, 2006). Specifically, it has been shown that adolescents with higher levels of BP characteristics scored consistently lower on life satisfaction over twenty subsequent years (Winograd, Cohen, & Chen, 2008).

Underlying Mechanisms: Rejection Sensitivity and Social Relations

One underlying mechanism of the association between BP characteristics and life satisfaction might be rejection sensitivity, defined as the degree to which an individual anxiously expects rejection from others (Levy, Ayduk, & Downey, 2001). Those with high levels of BP characteristics often interpret daily forms of rejection as a sign of personal disapproval or devaluation (Meyer, Ajchenbrenner, & Bowles, 2005; Staebler, Helbing, Rosenbach, & Renneberg, 2011). High rejection sensitivity, in turn, has consistently been linked to psychosocial problems such as rejection, depression, and loneliness, which are indicative of low levels of life satisfaction (McDonald, Bowker, Rubin, Laursen, & Duchene, 2010; Sandstrom, Cillessen, & Eisenhower, 2003). This suggests that rejection sensitivity may explain the negative relationship between BP characteristics and life satisfaction. However, rejection sensitivity might also act as a risk factor for low life satisfaction for those with high BP characteristics. BP characteristics longitudinally predicted lower satisfaction with the social network for adolescents with high rejection sensitivity, but not for those with low rejection sensitivity (Lazarus, Southward, & Cheavens, 2016), pointing to a moderating effect of rejection sensitivity.

Additionally, interactions between the environment and individual are crucial for the development of their personality (Belsky & Pluess, 2009). Both positive and negative aspects of social relations, such as social support and conflict, are related to development and health (Duffey & Somody, 2011). The individual’s social relations may be an underlying mechanism of the negative relationship between BP characteristics and life satisfaction. Individuals with high levels of BP characteristics seem to receive less social support than healthy individuals, report less positive relations overall, and experience more conflict in these relations (Belford, Kaehler, & Birrell, 2012; Daley, Burge, & Hammen, 2000). Likewise, BP characteristics have been negatively associated with social network quality one month later (Lazarus et al., 2016). Adolescents who received more social support, in turn, reported being more satisfied with their lives (Fife, Adegoke, McCoy, & Brewer, 2011). Alternatively, high levels of BP characteristics might be a risk factor for low life satisfaction for those with little social support and much conflict. This is in line with the buffering model (Cohen & Wills, 1985), which states that social support buffers and social conflict exacerbates the impact of stressful events on one’s life satisfaction. Indeed, when individuals with high levels of pathological personality traits are involved in positive social interactions this positively affects their adjustment (Paris, 2014). Similarly, social relations may influence the impact of BP characteristics on one’s life satisfaction.

The current study examined the direct effects and the mediating and moderating roles of rejection sensitivity and social relations in the association between BP characteristics and life satisfaction in a non-clinical group of young adolescents. It was expected that the influence of rejection sensitivity and social relations could (partly) account for the association between BP characteristics and life satisfaction, hence a mediating effect. Additionally, it was expected that low levels of rejection sensitivity and support from parents or friends would diminish and high levels of rejection sensitivity and conflict with parents or friends exacerbate any detrimental influence of BP characteristics on life satisfaction, hence a moderating effect.

Finally, findings concerning gender differences in prevalence and expression of BP characteristics have been inconsistent (Banzhaf et al., 2012; Bradley, Conklin, & Westen, 2005). More insight is needed into the (dis-)similarities between boys and girls regarding BP characteristics and the associated mechanisms and consequences. Explorative analyses were conducted to examine gender differences in the proposed mediation and moderation models.

Method

Sample

Participants were recruited from two high schools in a middle-sized city in the Netherlands, comprising three different education levels. 255 adolescents participated, including 144 boys (57%) and 111 girls (43% girls) between the ages of 15 and 19 (M_age = 16.27, SD = 0.70).

Measures

Satisfaction with life

The Dutch translation of the Satisfaction With Life Scale (SWLS) was used to measure life satisfaction (Desmyter, n.d.). The SWLS contains five items to measure global judgments of satisfaction with one’s life (e.g., ‘In most ways my life is close to my ideal.’). Items are scored on a 7-point Likert scale (1 = strongly disagree; 7 = strongly agree). The five items were summed. In the present study, Cronbach’s α was .81.

BP characteristics

The Dutch version of the Severity Indices of Personality Problems – Short Form (SIPP-SF) was used to measure BP characteristics (Andrea et al., 2007; de Viersprong, 2006). Participants were asked to answer 60 items and indicate to what extent each item was applicable to them during the last three months. Responses were given on a 4-point Likert Scale (1 = fully disagree; 4 = fully agree). The present study considered three higher order domains as indicative of BP characteristics: Self-control (e.g., ‘I lose control sometimes to the extent that people are frightened of me’), Identity Integration (e.g., ‘I am often confused about what kind of person I really am’) and Relational Functioning (e.g., ‘It is hard for me to get attached to someone else’), all containing 12 items (Samuel et al., 2012; Trull, Tomko, Brown, & Scheiderer, 2010). A total score on BP characteristics was computed by taking the mean of the scores on these three domains. Correlations of .64 and .50 (p < .01) were found between self-control and identity integration and relational functioning, respectively. Between identity integration and relational functioning this correlation was .67 (p < .01). For the total BPD scale, Cronbach’s α was .92.

Rejection sensitivity

To measure rejection sensitivity, participants completed the Dutch version of the Rejection Sensitivity Questionnaire (RSQ) (Downey & Feldman, 1996; Hanenberg, 2013). Participants were asked about their expectations of rejection in nine hypothetical situations (e.g., ‘You approach a close friend to talk to after having done or said something that seriously upset him/her.’). The responses on the items were given along two dimensions: (1) the anxiety and concern of being rejected and (2) the expectations of being accepted or rejected.. Items are rated on a 6-point Likert scale ((1) 1 = very unconcerned; 6 = very concerned; (2) 1 = very unlikely; 6 = very likely). A rejection sensitivity score for each situation was calculated by multiplying the level of rejection concern by the reverse of the level of acceptance expectancy. The total RSQ score was determined by the mean of the resulting nine scores. Cronbach’s α was .87.

Social network

To measure support and conflict in social relations, the Dutch translation of the Network of Relationships Inventory – Behavioral System Version (NRI-BSV) was used (Furman & Buhrmester, 2009; Van Aken & Hessels, 2012). Participants are asked to answer 24 items regarding their relationship with important persons in their social network (e.g., ‘How much do you seek out this person when you’re upset?’) on a 5-point Likert Scale (1 = little or none; 5 = the most). The NRI-BSV contains eight scales (e.g., ‘Seeks safe haven’; ‘Antagonism’) that load on two broad factors Support and Conflict. The mean scores for mother and father were combined into two factors: parent support and parent conflict. A factor analysis of the parent subscales indicated this two factor solution, parent support and conflict explained 54.42% and 26.94% of the total variance. The mean scores for friend-same-sex and friend-other-sex were combined into two factors: friend support and friend conflict. A factor analysis of the friend subscales indicated this two-factor solution, friend support and conflict explained 40.79% and 35.69% of the total variance. In the present study, Cronbach’s α was .82 for Parent Support, .74 for Friend Support, .72 for Parent Conflict, and .67 for Friend Conflict.

Procedure

Schools were recruited via invitation letters and two schools agreed to participate. The students received instructions on the procedure of completing the questionnaires and were informed that participation was voluntary and could be terminated at any moment. After giving informed consent, they completed the questionnaires in a computer room at their school. All questionnaires were completed online via a survey program.

Statistical analyses

Missing values were handled with Relative Mean Substitution (Raaijmakers, 1999). Twenty-two participants did not answer any items of a questionnaire and were excluded from the analyses, bringing the sample size down to 233 adolescents. Missing Value Analysis in SPSS revealed that adolescents with missing variables did not differ from adolescents with no missing variables (Little’s test > .05). Direct and mediation effects were tested with a model specifying direct and indirect regression paths in Mplus (Muthén & Muthén, 1998–2012). After centering variables and creating interaction terms, the moderation model was specified in Mplus by including independent variables and interaction terms.

Results

Descriptive statistics

Means, standard deviations, and gender differences regarding the study variables are shown in Table 1. Correlations between all study variables are displayed in Table 2.¹

Table 1. Means, standard deviations, and gender differences.

	M (SD) total	M (SD) boys	M (SD) girls	t (df)
BP characteristics	1.82 (.46)	1.86 (.43)	1.77 (.49)	1.51 (242)
Life satisfaction	4.80 (1.35)	4.78 (1.27)	4.81 (1.46)	−.16 (218.2)
Rejection sensitivity	7.79 (3.93)	8.29 (4.02)	7.11 (3.73)	2.35* (230.1)
Support parents	2.91 (.87)	2.79 (.89)	3.08 (.83)	−2.62* (242)
Conflict parents	2.06 (.70)	2.06 (.73)	2.05 (.66)	.03 (239)
Support friends	3.17 (.93)	2.90 (.94)	3.54 (.78)	−5.69** (235.7)
Conflict friends	1.61 (.51)	1.69 (.56)	1.50 (.41)	2.98* (236)

* p < .05

** p < .01.

Table 2. Correlations between all variables for boys (below diagonal) and girls (above diagonal).

	BP	LS	RS	PS	PC	FS	FC
BP	1	−.40**	.45**	−.29**	.39**	−.19	.22*
LS	−.30**	1	−.27**	.26**	−.38**	.09	−.03
RS	.42**	−.11	1	−.25*	.17	−.26**	.13
PS	−.32**	.25**	−.20*	1	−.36**	.30**	−.10
PC	.25**	−.11	.09	−.13	1	−.12	.30**
FS	−.25**	.21*	−.28**	.56**	.22*	1	.03
FC	.28**	−.08	.29**	.09	.40**	.15	1

Notes:

BP = borderline personality characteristics, RS = rejection sensitivity, LS = life satisfaction, PS = parent support, PC = parent conflict, FS = friend support, FC = friend conflict.

* p < .05

** p < .01.

Direct effect of rejection sensitivity and social relations

Looking at the direct effects, it was found that BP characteristics (B = −.81, SE = .22, β = −.27, p < .001) and conflict with parents (B = −.30, SE = .14, β = −.15, p = .028) were significant predictors of satisfaction with life. Support from parents (B = .14, SE = .12, β = .09, p = .221), support from friends (B = .09, SE = .11, β = .06, p = .427), conflict with friends (B = .18, SE = .18, β = .07, p = .321), and rejection sensitivity (B = −.001, SE = .02, β = −.002, p = .979) were not significant predictors. Looking at boys and girls separately using multi-group testing, it appeared that BP characteristics was a significant predictor of satisfaction with life for both boys and girls (Boys: B = −.69, SE = .28, β = −.23, p = .014; Girls B = −.83, SE = .33, β = −.25, p = .011). For boys, however, conflict with parents did not predict satisfaction with life (B = −.12, SE = .17, β = −.07, p = .482), whereas for girls it did (B = −.61, SE = .22, β = −.28, p = .006).

Mediation effects of rejection sensitivity and social relations

It was found that conflict with parents mediated the association between BP characteristics and satisfaction with life. This resulted in a significant overall indirect effect (B = −.25, SE = .12, β = −.08, p = .044). Rejection sensitivity, parent support, friend support, and friend conflict did not mediate the association between BP characteristics and satisfaction with life (see Table 3). When considering gender differences in this indirect effect, however, it appeared that conflict with parents only mediated this association for girls (B = −.33, SE = .14, β = −.11, p = .019), but not for boys (B = −.05, SE = .07, β = −.02, p = .467).

Table 3. Results of the mediation and moderation analyses (maximum likelihood estimation).

		B	SE	β	p
Mediation	Rejection sensitivity	−.03	.09	−.01	.734
	Parent support	−.10	.07	−.03	.167
	Parent conflict	−.14	.07	−.05	.040
	Friend support	−.04	.05	−.01	.517
	Friend conflict	.06	.06	.02	.309
Moderation	BPC × rejection sensitivity	.08	.06	.06	.138
	BPC × parent support	.22	.29	−.06	.454
	BPC × parent conflict	−.26	.34	−.03	.443
	BPC × friend support	−.10	.26	.00	.711
	BPC × friend conflict	.00	.43	.10	1.00

Notes:

Bs reported for mediation analyses are the estimates of the indirect effect when rejection sensitivity or support and conflict variables are modeled as mediator.

Based on modification indices correlations between support and conflict variables were included in the mediation model. Model fit statistics for this model indicated adequate fit (χ² = 29.94, df = 6, p < .001, CFI = .914, SRMR = .058). Model fit statistics for the moderation model could not be assessed as this model is saturated (Muthén & Muthén, 1998–2012).

Moderation effects of rejection sensitivity and social relations

It was found that none of the investigated variables moderated the association between BP characteristics and satisfaction with life (see Table 3). However, when considering gender differences, it appeared that rejection sensitivity did moderate this association, but only for girls (B = .21, SE = .08, β = .26, p = .010) and not for boys (B = −.001, SE = .08, β = −.001, p = .989). See Figure 1.

Figure 1. (a) Moderating effects of rejection sensitivity on the relationship between BP characteristics and life satisfaction for boys. (b) Moderating effects of rejection sensitivity on the relationship between BP characteristics and life satisfaction for girls.

Note: Low and high BP characteristics and rejection sensitivity reflect 1 SD below and above the sample mean.

Discussion

The aim of this study was to examine the association between personality characteristics that are indicative of BPD and adolescents’ satisfaction with life. Additionally, rejection sensitivity and social relations were expected to be related to BPD such that differing levels could lead to important consequences for personal functioning and satisfaction. A deepened understanding of these associations could provide suggestions on how to influence the life satisfaction of non-clinical adolescents with pathological personality characteristics. Our results showed, in congruence with our expectations, that BP characteristics were negatively associated with life satisfaction. Furthermore, support in relations was positively and conflict in relations and rejection sensitivity were negatively associated with life satisfaction. Strikingly, BP characteristics were negatively associated with support from friends for boys, but not for girls. This link between BP characteristics and friendships has not often been studied, and rarely in relation to gender differences. Our results therefore point in interesting directions for further research. King and Terrance (2006) found that borderline personality traits in college students were associated with viewing closest friends as unsafe and untrustworthy, more feelings of uncomfortableness during interaction, and less friendship security for both boys and girls. An explanation for our findings could be that boys who exhibit BP characteristics may have a more aggressive, disruptive and antisocial pattern (Bradley et al., 2005). As a result, they may experience less support from friends than girls (Fuhrer & Stansfeld, 2002). Indeed we found that boys scored higher than girls on rejection sensitivity and conflict with friends. Boys may be more likely than girls to experience direct aggression and rejection, resulting in the anticipation of negative social interactions and possible direct rejection by their friends (London, Downey, Bonica, & Paltin, 2007). For girls, this process may be more indirect and therefore harder to detect. Girls did score higher than boys on support from both parents and friends. This is in line with previous findings suggesting that girls mention more close relations, report greater satisfaction with these relations and are more likely to seek social support as a coping strategy (Eschenbeck, Kohlmann, & Lohaus, 2007; Fuhrer & Stansfeld, 2002). Nevertheless, we did find that both girls and boys with higher levels of BP characteristics experience more conflict in friendships.

Influence of rejection sensitivity and social relationships (mediation)

Conflict with parents was found to mediate the association between borderline personality characteristics and satisfaction with life, however, only for girls. A possible explanation may be that for girls, in line with the correlations found in this study, relationships with parents are more important for determining life satisfaction than for boys. This is supported by Cheng and Chan (2006), who found that social relationships are a stronger determinant of life satisfaction in women than in men, indicating that girls may attach greater value to supportive relationships with their parents and are more negatively affected by conflict in these relationships than boys (Eschenbeck et al., 2007). Relationships with friends could be less influential than relationships with parents because of the difficulties adolescents, who score high on BP characteristics, experience in establishing long-lasting and stable friendships (Johnson, Chen, & Cohen, 2004). Additionally, even though friends become more important during adolescence, relationships with parents remain the best predictor of life satisfaction and emotional problems (Nickerson & Nagle, 2004).

Contrary to our expectations, no mediating effects of support in social relations were found. The present study used a non-clinical sample and it may be expected that these adolescents generally have a supportive relationship with their parents, in which conflict is a more salient event and thus may cause a stronger effect than support. Additionally, friendships may be considered as inherently supportive relationships, limiting the effect of friend support. The presence of a good friend in and of itself implies the presence of a supportive friendship (Hartup, 1996). Moreover, no mediating effects of rejection sensitivity were found. This could be explained by the fact that our participants generally demonstrated normal levels of rejection sensitivity. These associations are expected to be found in a clinical sample (Staebler et al., 2011) and future research should examine these effects.

Diminishing or exacerbating effects of rejection sensitivity and social relations (moderation)

It was found that high levels of rejection sensitivity exacerbate the negative effect of BP characteristics on life satisfaction, again only in girls. This could be explained by the process that high rejection sensitivity leads to a self-fulfilling prophecy in which girls’ expectations of rejection lead them to engage in defensive actions against or withdrawal from others, increasing the likelihood of actual rejection, an increase in social conflict (Wang, McDonald, Rubin, & Laursen, 2012) and, in turn, a decrease in life satisfaction. Higher levels of rejection sensitivity have also been linked to temporary reductions in self-concept clarity (Ayduk, Gyurak, & Luerssen, 2009). In turn, a clear concept of the self has been linked to positive self-esteem and psychological well-being (Usborne & Taylor, 2010). This might act as an underlying mechanism, which future research could examine.

The expectations regarding social support from and conflict with parents and friends as a moderator were not supported. Considering the absence of direct effects of support from parents and friends and conflict with friends on life satisfaction, as well as the non-clinical levels of borderline personality traits in our sample, the effects were likely not strong enough to detect moderation. However, considering that borderline personality traits were related to support and conflict in relations, future research looking at clinical samples might be able to detect these effects.

Limitations and recommendations

The use of the SIPP-SF as a measurement of BP characteristics can be seen as a strength and a limitation of the present study. SIPP-SF scores were found to be normally distributed throughout the current sample, supporting the dimensional view regarding BP characteristics. However, this dimensional approach makes it difficult to compare the current results with results of studies that used a categorical approach. Additionally, since the SIPP-SF was developed to identify the core components of maladaptive personality functioning, instead of specifically measuring BP characteristics, no scoring manual exists for computing the degree of BP characteristics. Correlations and the reliability of the total BPD scale suggested that self-control, relational functioning, and identity integration measure the same construct, i.e., BP characteristics. However, more research on using these subscales is warranted to validate the use of the SIPP-SF as an adequate measure to identify BP characteristics.

As with all cross-sectional studies, no inferences can be made about causation in the investigated relationships. The present study assumed that BP characteristics, being more stable characteristics of an individual throughout the lifetime, predicted the amount of rejection sensitivity and life-satisfaction, as more instable or fluctuating factors.

Implications

The differing results for support and conflict within the social relationships indicate that these two aspects may have a differential influence on the investigated factors. The results lead us to suggest that conflict resolution and management might be an important issue that should be focused on within the educational system, to help adolescents, especially girls, become better at resolving conflicts. Furthermore, our findings show that the role of parents should not be disregarded for adolescents. The investigated age group could be an explanation for the finding that conflict with parents plays such an important role, considering that adolescence is a period characterized by increased conflict with parents (Arnett, 1999). Future research could also examine the differential effect of rejection sensitivity for boys and girls, as higher levels of rejection sensitivity exacerbated the effect of BP characteristics for girls, while boys actually scored higher on rejection sensitivity.

Further, based on the results, it is strongly recommended to differentiate between support and conflict in the treatment of patients with BPD. In therapy it is important not only to ensure that patients have an adequate support network, but also to resolve the conflictuous relationship they might have with their parents. It may be beneficial to emphasize that even though a relationship is predominantly conflictuous, support can still be derived from it, and similarly, a highly supportive relationship may be disturbed by a single conflictuous event (Arnett, 1999). This differentiation may also be made in various social skills trainings, focusing on establishing and maintaining social relationships as well as conflict resolution and management.

In conclusion, the current study adds to the literature by showing that rejection sensitivity and social relationships have an impact on the associations between BP characteristics and life satisfaction, and this differs for boys and girls. Although social support does not act as a buffer for boys or girls, being sensitive to rejection and conflict with parents do act as pitfalls for girls. These factors explain or strengthen the association between BP characteristics, which can also be found in a non-clinical population, and adolescent’s satisfaction with life. Therefore, it is advised to focus on these aspects in a school curriculum, especially with girl students.

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

¹ To gain an indication of the severity of the BP characteristics, the current sample was compared to a normative non-clinical sample and T scores were computed (de Viersprong, 2006). T-scores below 40 indicate impaired adaptive functioning, and below 30 indicate severely impaired functioning. The current non-clinical sample demonstrated adaptive functioning as indicated by mean T-scores above 40. However, there were adolescents who indicated clinical levels of BP characteristics as demonstrated by the range of T scores (self-control: M_T= 45.54, SD_T = 10.02, range_T = 11.44–61.34; identity integration: M_T = 48.26, SD_T = 9.28, range_T = 14.91–59.09; relational functioning: M_T = 47.35, SD_T = 7.64, range_T = 29.15–62.98) (de Viersprong, 2006).

The severity of rejection sensitivity scores was examined through comparison to a normative non-clinical sample (M = 9.66, SD = 3.03, range = 2.40–23.50) (Downey & Feldman, 1996). Again, adolescents in the current sample displayed normal levels of rejection sensitivity on average, though there were also adolescents who demonstrated extremely high levels (M = 7.79, SD = 3.93, range = 1–18).