- Email: [email protected]
Do borderline personality disorder features and rejection sensitivity predict social network outcomes over time?
PAID-07326; No of Pages 6 Personality and Individual Differences xxx (2016) xxx–xxx
Contents lists available at ScienceDirect
Personality and Individual Differences journal homepage: www.elsevier.com/locate/paid
Do borderline personality disorder features and rejection sensitivity predict social network outcomes over time? Sophie A. Lazarus ⁎, Matthew W. Southward, Jennifer S. Cheavens Department of Psychology, Ohio State University, United States
a r t i c l e
i n f o
Article history: Received 20 June 2015 Received in revised form 5 November 2015 Accepted 12 February 2016 Available online xxxx Keywords: Borderline personality disorder Social networks Interpersonal functioning Rejection sensitivity Social support
a b s t r a c t Social functioning is routinely understood to be disrupted for those with BPD features; however there is little understanding of how BPD features and BPD-relevant traits impact social network characteristics over time. We hypothesized that BPD features negatively predict social network quality and composition and that rejection sensitivity (RS) would affect these relations. To examine this, a sample of female college students (N = 127) was recruited and followed over one month. BPD features predicted lower ratings of social network quality and aspects of network composition. BPD features exerted an indirect effect through one-week RS on perceived levels of conflict and criticism as well as on number of partners in the network at one-month follow-up. Moderation analyses revealed that BPD features predicted lower social network satisfaction and support at one month for those with high RS, but did not impact satisfaction or support for those with low RS. These results indicate that even non-clinical levels of BPD psychopathology are related to poor social network outcomes. These findings also highlight RS as a potential mechanism by which BPD features predict lower social support and satisfaction and a potential risk factor for higher conflict and criticism within social networks. © 2016 Elsevier Ltd. All rights reserved.
1. Introduction Theoretical and empirical accounts point to difficult interpersonal relationships as a major source of distress for those with borderline personality disorder (BPD; Gunderson, 2007; Hilsenroth, Menaker, Peters, & Pincus, 2007). There is increasing evidence that BPD features are related to both objective (e.g., composition of networks) and subjective (e.g., ratings of network quality) social network constructs. In terms of objective social network outcomes, BPD status is related to smaller social networks, measured by fewer interaction partners per day (Stepp, Pilkonis, Yaggi, Morse, & Feske, 2009), and more BPD features predict the generation of fewer partners available to fulfill the need for social support (Zielinski & Veilleux, 2014). BPD criteria are also related to an increased proportion of conflictual or romantic partners in one's social network (e.g., Clifton, Pilkonis, & McCarty, 2007; Daley, Burge, & Hammen, 2000). Consistent with these compositional differences, BPD features have been associated with less satisfaction and support from romantic partners (Bouchard, Sabourin, Lussier, & Villeneuve, 2009; Daley et al., 2000) and from social partners in general (Zielinski & Veilleux, 2014). However, there is increasing evidence that positive
⁎ Corresponding author at: 181, Psychology Building, 1835 Neil Avenue, Columbus, OH 3210, United States. E-mail address: [email protected] (S.A. Lazarus).
social relationships may result in lower levels of anger, a BPD criterion (Kuhlken, Robertson, Benson, & Nelson-Gray, 2013), and that marriage may positively predict overall global functioning and symptom status over time for those with BPD features (Links & Heslegrave, 2000). Thus, given the protective nature of stable, satisfying relationships and the likelihood that BPD features may interfere with forming or keeping such relationships, it is imperative to determine what predicts high quality relationships at high levels of BPD features. A growing body of research indicates that problematic social network outcomes are associated with BPD features (e.g., Daley et al., 2000; Zielinski & Veilleux, 2014), even in samples not diagnosed with or selected for BPD status. Understanding social network dysfunction across the continuum of BPD features is important given evidence that even minimal levels of BPD pathology have clinical significance (Zimmerman, Chelminski, Young, Dalrymple, & Martinez, 2012) and subclinical BPD features are longitudinally associated with difficulties in academic achievement, mood, and interpersonal functioning (Trull, Useda, Conforti, & Doan, 1997). To date, most existing research has examined the effects of BPD features on social network characteristics without considering the impact of vulnerabilities associated with BPD. Examining specific personality characteristics that may put those with BPD features at increased risk for poor social network outcomes or contribute to specific social network characteristics over time can give us a more precise understanding of the interplay between BPD features and disrupted social networks.
http://dx.doi.org/10.1016/j.paid.2016.02.032 0191-8869/© 2016 Elsevier Ltd. All rights reserved.
Please cite this article as: Lazarus, S.A., et al., Do borderline personality disorder features and rejection sensitivity predict social network outcomes over time?, Personality and Individual Differences (2016), http://dx.doi.org/10.1016/j.paid.2016.02.032
2
S.A. Lazarus et al. / Personality and Individual Differences xxx (2016) xxx–xxx
1.1. Rejection sensitivity (RS) and BPD
1.2. Current study
RS, defined as the tendency to anxiously expect, readily perceive, and react strongly to perceived rejection (Downey & Feldman, 1996), frequently characterizes BPD, yet the two constructs are not synonymous. The empirical association between BPD features and RS has been demonstrated in several studies (e.g., Berenson, Downey, Rafaeli, Coifman, & Paquin, 2011; Tragesser, Lippman, Trull, & Barrett, 2008) and some researchers have investigated RS as a mechanism through which BPD results in problematic outcomes. Among those diagnosed with BPD, Selby, Ward, and Joiner (2010) found that BPD symptoms predicted RS, which predicted dysregulated emotions and eating behavior. In undergraduate samples, RS mediated the relation between BPD features and facial trust appraisal (Miano, Fertuck, Arntz, & Stanley, 2013) and between BPD features and number of social contacts (Zielinski & Veilleux, 2014). However, Zielinski and Veilleux (2014) found that RS did not mediate the relation between BPD features and social support satisfaction. These results can be interpreted to suggest that elevated fears of social rejection in BPD may decrease regulatory ability, leading to impulsive behaviors, less trust of others, and fewer social contacts but that the relation between BPD and satisfaction with one's social support may be a direct effect, not dependent on RS. While the association between BPD and RS across investigations suggests that RS may be important in understanding how and when BPD features predict social network characteristics, these meditational studies were all cross-sectional, limiting our understanding of temporal links. Assessing the composition and quality of social networks can be a daunting task given the number of decisions that must be made to appropriately operationalize and define the construct. One of the first decisions is whether the social network should be examined at one time point or over time. The majority of compositional research to date has examined social network differences between those with BPD and various control participants either at one time point to assess how BPD features and social network qualities are concurrently related (e.g., Clifton et al. (2007)) or over relatively short time periods (e.g., one week; Stepp et al., 2009) to determine how particular interactions impact short-term outcomes such as mood. However, these designs do not permit an investigation of how social networks are affected by BPD features over time. Second, who is considered to be in one's social network is an important aspect of social network research. Some researchers have assessed specific relationships (e.g., romantic partners; Bouchard et al., 2009) while others have examined partners who might be available in particular contexts (e.g., Zielinski & Veilleux, 2014). While these designs are informative, it is also important to assess social networks more broadly, as those with whom the individual frequently interacts may impact their emotional functioning and available support.
We examined how BPD features impacted social network characteristics using ratings of social network relationships over time. Existing research provides some insight into specific characteristics of the social networks of those with BPD and BPD features using various methods, such as daily diary assessment and cross-sectional self-report. However, this research is limited and there is no consensus on the best methods to assess social networks. Given this, we developed a measure, described below, to assess relationships with all partners with whom individuals frequently interact. Based on previous findings, we assessed composition of the social networks (e.g., number of total partners, romantic partners, and partners to whom one had stopped speaking) as well as the quality of each relationship (i.e., satisfaction, support, conflict, and criticism). We tested whether BPD features predicted social network characteristics one month later. We also assessed the influence of RS on these associations. While several studies have examined RS as a mediator of the relations between BPD and socially relevant outcomes (e.g., Miano et al., 2013; Zielinski & Veilleux, 2014), RS also has been tested as a moderator that amplifies risk for deficits in social functioning at higher levels of BPD features (Gardner, Qualter, Stylianou, & Robinson, 2010). We tested RS as both a moderator and a mediator of the relations between BPD features and social network outcomes to clarify whether RS is the mechanism through which these relations exist or a risk factor for poor social outcomes for those with BPD features. Examining interpersonal relationships in this manner allows for a more complete picture of social networks and BPD features and a better understanding of potential risk factors and mechanisms for disrupted network quality. 2. Materials and methods 2.1. Participants and procedure This study was open to all female introductory psychology students. Because BPD is primarily diagnosed in women (American Psychiatric Association, 2013) and there are tend to be differences in social networks (Kendler, Myers, & Prescott, 2005) and RS (Downey, Freitas, Michaelis, & Khouri, 1998) based on gender, we limited participation to female students. At baseline, participants completed consent forms and questionnaires assessing personality variables (i.e., BPD features, RS) and social network characteristics. One week and one month after baseline, participants were emailed a link to complete the same questionnaires. Participants were compensated with either partial course credit or online gift cards (because the semester ended prior to study completion for some participants). The sample consisted of 127 female students at a large Midwestern university. Participants ranged in age from 18 to 32 (M = 19.57,
Table 1 Baseline means, standard deviations, and correlations. Measure a
1. PAI-BOR 2. RSQb 3. Total partners 4. Romantic partners 5. Cut-off partners 6. Satisfaction 7. Support 8. Conflict 9. Criticism
M
SD
21.02 8.86 11.17 1.09 .95 2.42 2.38 1.38 .86
10.07 3.32 6.29 1.94 1.41 .39 .47 .52 .52
1
2
3
4
– .36⁎⁎ .25⁎⁎ −.24⁎⁎ −.23⁎ −.31⁎⁎ −.26⁎⁎
– .23⁎⁎ −.16 −.10 −.09 −.01
5
6
7
8
–
.26⁎⁎ .04 .20⁎ .35⁎⁎ −.42⁎⁎ −.25⁎⁎ .34⁎⁎ .22⁎
– −.14 −.07 .18⁎ −.28⁎⁎ −.16 .40⁎⁎ .29⁎⁎
– −.37⁎⁎ −.33⁎⁎ .08 .01
– −.58⁎⁎ −.21⁎ −.06
– .05 .05
–
.46⁎⁎
⁎ p b .05. ⁎⁎ p b .01. a Personality Assessment Inventory — Borderline subscale. b Rejection Sensitivity Questionnaire.
Please cite this article as: Lazarus, S.A., et al., Do borderline personality disorder features and rejection sensitivity predict social network outcomes over time?, Personality and Individual Differences (2016), http://dx.doi.org/10.1016/j.paid.2016.02.032
S.A. Lazarus et al. / Personality and Individual Differences xxx (2016) xxx–xxx
SD = 2.50) years old and over half were Caucasian (66.1%). The majority of participants were single (66.1%) with a large minority reporting they were in non-cohabitating relationships (27.6%). There were no exclusion criteria. Inclusion criteria were that participants be female, 18 years old or older, and enrolled in an introductory psychology course.
3
Table 2 Baseline PAI-BOR scores predicting one-month SNA outcomes. Poisson regression
df
b
SE
β
χ2
Pseudo R2a
Number of partners Number of romantic partners Number of cut-off partners
1, 125 1, 125 1, 125
.01 .03 .04
b .01 .01 .01
.06 .31 .36
3.71 15.69 12.03
.01 .07 .06
2.2. Study measures 2.2.1. Borderline features The Personality Assessment Inventory — Borderline subscale (PAIBOR; Morey, 1991) is a 24-item, self-report inventory designed to assess BPD features in adults. In this sample, it demonstrated good internal consistency (Cronbach's α = .85) and good test–retest reliability over four weeks (r = .76). 2.2.2. Rejection sensitivity The Rejection Sensitivity Questionnaire (RSQ; Downey & Feldman, 1996) consists of 18 scenarios in which participants imagine requesting something from another person that makes them vulnerable to rejection. Participants rate how anxious they are that the person will respond negatively to the request on a 1 (very unconcerned) to 6 (very concerned) scale, and the likelihood that the person will respond positively to the request on a 1 (very unlikely) to 6 (very likely) scale. Total scores are calculated by taking the inverse of the likelihood ratings, multiplying by the anxiety ratings, and taking the average, so that high scores reflect high anxiety and high likelihood of rejection. In this sample, the scale demonstrated good internal consistency (Cronbach's α = .88) and test–retest reliability over four weeks (r = .81). 2.2.3. Social network Created for this study, the Social Network Assessment (SNA) prompts participants to list people with whom they have interacted two or more times for at least 5 min each time in the previous week. Participants listed up to 24 partners in order of the frequency of interaction. For each partner listed, participants rated their satisfaction, support, conflict, and criticism in the relationship on a 1 (not at all) to 4 (very much) Likert scale. Ratings were averaged for each participant at each time point to produce estimates of global network satisfaction, support, conflict, and criticism. Participants identified whether each partner was a former or current romantic partner (i.e., romantic partner) and if they had “cut off” speaking to each partner in the past month. 2.3. Analytic strategy We first determined the baseline and one-month characteristics of our sample. Poisson regression analyses were conducted to test whether baseline BPD features predicted one-month social network composition count variables (i.e., total number of partners, number of romantic partners, and number of cut-off partners). We conducted OLS linear regression analyses to test whether baseline BPD features predicted onemonth social network quality variables (i.e., satisfaction, support, conflict, and criticism). Mediation analyses were conducted using the PROCESS macro (Hayes, 2013) with 10,000 percentile bootstraps to test whether baseline BPD features exerted an indirect effect on onemonth social network outcomes through one-week RS. Finally, we entered BPD, RS, and their product as predictors of social network outcomes to test whether RS moderated the impact of BPD features on one-month social network characteristics. All analyses were conducted in SPSS version 20. 3. Results 3.1. Baseline characteristics At baseline, participants reported PAI-BOR scores that were consistent with other undergraduate samples (T-score = 48; Morey, 1991)
OLS regression
df
b
SE
β
F
R2
Satisfaction Support Conflict Criticism
1, 125 1, 125 1, 125 1, 125
−.01 −.01 .02 .01
.01 .01 .01 .01
−.26 −.21 .26 .22
8.88 5.58 9.14 6.52
.07 .04 .07 .05
Note. PAI-BOR = Personality Assessment Inventory — Borderline subscale. SNA = Social Network Assessment. Each line represents a separate model. deviance ðPAIBORÞ a Pseudo R2, calculated as 1 deviance . ðintercept onlyÞ
with 6% of the sample scoring above 37 on the PAI-BOR, suggesting clinically relevant BPD features. RSQ scores were also comparable to those found in other undergraduate samples (Downey & Feldman, 1996; Zielinski & Veilleux, 2014). Table 1 includes the means, standard deviations, and correlations among variables at baseline. 3.2. Do baseline BPD features predict one-month social network outcomes? BPD features at baseline predicted interacting with more current or former romantic partners (p b .01), cutting off more partners (p b .01), and marginally predicted more total partners in the network (p = .05) one month later (Table 2). BPD features also predicted more conflict (p b .01) and criticism (p = .01) as well as less satisfaction (p b .01) and support (p = .02) one month later.1,2 3.3. Does RS mediate the impact of BPD on one-month social network outcomes? Next, we attempted to extend Zielinski and Veilleux's (2014) crosssectional findings by testing whether baseline BPD features exerted an indirect effect through one-week RS on one-month social network outcomes. BPD features exerted an indirect effect through RS on the total number of partners (Table 3). BPD features predicted greater RS at one week, which predicted fewer total partners at one month. BPD features did not exert an indirect effect through RS on the number of current or former romantic partners or the number of cut-off partners. Like Zielinski and Veilleux (2014), we found no evidence of mediation for one-month satisfaction or support. However, there was an indirect effect of BPD features through RS on both conflict and criticism at one month. Baseline BPD features predicted greater RS at one week, which predicted greater average perceived conflict and criticism at one month.3 3.4. Does RS moderate the impact of BPD on one-month social network outcomes? Finally, we examined whether RS moderated the impact of BPD features on social network outcomes. The product of baseline BPD features 1 We ran each of the above models controlling for the baseline measure of the respective social network characteristic to test whether BPD features predicted change in each social network outcome. BPD features only predicted one-month change in the number of cut-off partners (χ2(1 N = 128) = 4.92, B = 0.02, p = .03; all other ps N .06). 2 When examined at baseline only, all findings from the Poisson and OLS regression models are interpreted equivalently to the one-month findings (i.e., the signs of the beta weights are in the same direction and result in the same inferential conclusions regarding statistical significance). 3 We ran each of the above models controlling for baseline measures of the respective social network characteristic to test whether BPD features exerted an indirect effect on changes in each social network outcome through RS. BPD features only exerted an indirect effect through RS on changes in criticism (ab = .002, SE = .002, 95% CI: .000, .006).
Please cite this article as: Lazarus, S.A., et al., Do borderline personality disorder features and rejection sensitivity predict social network outcomes over time?, Personality and Individual Differences (2016), http://dx.doi.org/10.1016/j.paid.2016.02.032
4
S.A. Lazarus et al. / Personality and Individual Differences xxx (2016) xxx–xxx
Table 3 One-week RSQ scores mediating baseline PAI-BOR and one-month SNA outcomes. Outcome
ab
SE
LLCI (95%)
ULCI (95%)
Number of partnersa Number of romantic partners Number of cut-off partners Satisfaction Support Conflict Criticism
−.030 b.001 b.001 b −.001 b −.001 .004 .005
.018 .004 .002 .001 .001 .002 .002
−.071 −.008 −.004 −.003 −.003 .001 .001
−.001 .009 .006 .003 .003 .008 .010
Note. The indirect effect (ab) is significantly different from zero if the 95% confidence interval between the lower limit (LLCI) and upper limit (ULCI) does not include zero. RSQ = Rejection Sensitivity Questionnaire. PAI-BOR = Personality Assessment Inventory — Borderline subscale. SNA = Social Network Assessment. ab = −.03, SE = .02, 95% CI: –.060, .003. a Not significantly different from zero when using baseline variables only.
and RS predicted total number of partners at one month (χ2(1 N = 128) = 4.98, b = .002, p = .03). Specifically, BPD features more strongly predicted the total number of partners when RS was high than when RS was low (Fig. 1). However, people low in RS reported more partners than those high in RS, regardless of BPD features. RS did not moderate the impact of BPD features on the number of current or former romantic partners (χ2(1 N = 128) = 3.51, b = .004, p = .06) or cut-off partners (χ2(1 N = 128) = 0.40, b = −.002, p = .53).4 The product of BPD features and RS also predicted one-month social network satisfaction (F(1, 124) = 4.62, b = −.003, p = .03) and support (F(1, 124) = 5.54, b = − .004, p = .02) but not conflict (F(1, 124) = 1.46, b = − .002, p = .23) or criticism (F(1, 124) = 3.67, b = −.003, p = .06). Specifically, BPD features predicted lower social network satisfaction and support at one month for those with high RS, but did not impact satisfaction or support for those with low RS (Fig. 2).5,6
4. Discussion Taken together, these findings help characterize the relations between BPD features and social network outcomes and have implications for understanding the context in which interpersonal difficulties occur for those with heightened features of BPD. First, the results qualify the effect of BPD features on perceived social network quality. Although BPD features negatively predicted both perceived support from and satisfaction with one's social network, this effect was strongest at higher levels of RS. That is, the tendency to anxiously expect and perceive rejection may amplify fears of abandonment, impulsivity, and anger experienced at high levels of BPD features, ultimately eroding perceptions of social support and satisfaction. In a longitudinal study of older adolescents, Marston, Hare, and Allen (2010) found that those high in RS were rated by close friends as less socially accepted and less close one year later. Thus, it is possible that RS impacts social networks from “both ends” by impacting the individual's sense of satisfaction and security in relationships as well as the degree to which others feel close and connected to that individual. Our results suggest that the combination of BPD features and RS may predict perceived deficits in social satisfaction and support as quickly as one month later. 4 Using only baseline variables, the product of BPD features and RS predicted the number of cut-off partners, B = −.006, SE = .003, p = .04. The interpretations of total number of partners and number of romantic partners remained the same using baseline variables. 5 We ran each of the above models controlling for baseline measures of the respective social network characteristic to test whether RS moderated the effect of BPD features on changes in social network outcomes. The interaction of RS and BPD significantly predicted changes in the total number of partners (χ2(1 N = 128) = 23.43, B = .004, p b .01), satisfaction (F(1, 123) = 8.37, B = − .004, p b .01), and support (F(1, 123) = 7.18, B = −.003, p b .01). These effects were in the same direction as the predictive tests. 6 Entering only baseline variables, the product of BPD features and RS did not predict social network satisfaction or support (ps N .30). The interpretations of conflict and criticism remained the same using baseline variables.
Fig. 1. RS moderates the impact of BPD features on the total number of social network partners one month later.
BPD features also positively predicted average conflict and criticism in one's network one month later, and RS mediated this effect. That is, BPD features predicted increased sensitivity to rejection one week later, and RS, in turn, predicted greater perceived conflict and criticism in one's network at the one-month assessment. These findings build on the results of Downey et al. (1998), that high-RS women engage in more negative relationship behaviors (e.g., verbal put-downs, adopting a negative tone of voice, denying responsibility) after conflicts with their partners, leading to relationship erosion. Our findings suggest that women with higher BPD features perceive more conflict and criticism in their network because they are more sensitive to cues of rejection. Future researchers should investigate whether conflict behaviors jointly mediate this relation. Together, these findings suggest that RS may be a mechanism by which BPD features impact perceptions of conflict and criticism as well as a vulnerability that interacts with BPD features leading to perceptions of less satisfaction and support from social networks. Finally, we found that BPD features marginally predicted more social network partners overall and that RS both moderated and mediated this effect in separate models. In the mediation model, BPD features exerted a negative indirect effect on the total number of partners through RS. BPD features predicted greater RS one week later, which predicted fewer partners one month later. In the moderation model, BPD features positively predicted the number of partners at one month when RS was high but were unrelated to the number of partners when RS was low. Notably, those lower in RS tended to report more partners at all levels of BPD features than those higher in RS. These results suggest that among people with the same level of RS, those with more BPD features tend to report more social network partners. It is possible that other aspects of BPD such as impulsivity or novelty seeking, which were not the focus of this study, also contribute to pursuing a greater number of social partners.
Fig. 2. RS moderates the impact of BPD features on average social network satisfaction one month later.
Please cite this article as: Lazarus, S.A., et al., Do borderline personality disorder features and rejection sensitivity predict social network outcomes over time?, Personality and Individual Differences (2016), http://dx.doi.org/10.1016/j.paid.2016.02.032
S.A. Lazarus et al. / Personality and Individual Differences xxx (2016) xxx–xxx
The findings from this study are consistent with previous research showing that BPD features are related to lower relationship satisfaction and more high-conflict relationships (Clifton et al., 2007; Zielinski & Veilleux, 2014). While previous researchers have found evidence of decreased satisfaction in several specific interpersonal relationships (Bouchard et al., 2009; Stepp et al., 2009; Zielinski & Veilleux, 2014), a significant strength of the current study is the assessment of a wide, unselected range of partners over time. Assessing social network partners in this way may also explain why we found that BPD features were positively related to the total number of partners while other researchers found that BPD was related to fewer supportive partners (Zielinski & Veilleux, 2014) and fewer social interactions (Stepp et al., 2009). BPD features may actually be related to interacting with more partners, but perhaps also to the perception of those partners as less supportive or reliable for daily needs. This is consistent with research showing that BPD features are related to a preference for novel over familiar partners in a behavioral laboratory task (Cheavens, Lazarus, & Herr, 2014). The presence of more interaction partners in the network, or the behavior of seeking out new partners, may interfere with the ability to invest in individual relationships. It is important to consider details such as specific methodology for collecting network information when considering different findings regarding social network qualities. These results also establish the relation between BPD features and perceived conflict and criticism in an undergraduate sample. These findings build on those of Russell, Moskowitz, Zuroff, Sookman, and Paris (2007) who found that individuals with BPD reported engaging in more quarrelsome behaviors over 20 days than healthy controls. The significance of this finding in a non-clinical sample suggests that even elevated features of BPD among those with normal variation in symptoms of the disorder are related to networks with higher levels of conflict and criticism and reports of quarrelsome behavior are consistent with reports of conflictual relationships. Higher BPD features were also associated with more network members who were current or former romantic partners and more partners who were cut off in the past month. The presence of more conflictual partners may contribute to greater perceived conflict and criticism among network partners. However, while RS mediated the relation between BPD features and network conflict and criticism, it did not mediate or moderate the relations between BPD features and romantic or cut-off partners. Anxiously expecting and perceiving rejection may more strongly influence perceptions of partners than the type of partners in a person's network. Some aspects of the current study limit the strength of the conclusions we can draw and point to areas for future research. First, while it is important to establish the relations between BPD features, RS, and various social network outcomes in those with normal variation in these features, these relations may be different for individuals diagnosed with BPD. Perhaps in a clinical sample, where the relation between BPD and RS is likely stronger, our ability to detect the role of RS on social networks may be limited. However, the relation between BPD and RS found in clinical samples (e.g., Selby et al., 2010) suggests that these constructs are far from redundant, and this should be examined empirically. In addition, RS is likely not the only personality characteristic that increases the risk of poor social network outcomes for those with BPD features. Affective instability, interpersonal aggression, or emotion regulation difficulties may also interact with BPD diagnosis or features to impact the quality or composition of social networks. Examining other potential vulnerabilities in those with BPD features would contribute to our understanding of the processes by which network differences develop or are maintained. In addition, although information on specific social network partners was obtained at baseline and one month, we cannot assume that the same partners define each participant's network at each time point. Thus, while we report the average social network quality and composition for each participant, we can only say that change over time is due to global social network changes, not changes in specific relationships. In future studies, it would be useful to assess the same partners at each
5
assessment to evaluate change within a given relationship as well as on average in the network. Finally, while the longitudinal design is a strength of this study, longer follow-up periods may reveal more variance in social network outcomes and increase the opportunity to determine how changes in social networks are related to BPD features. This study contributes to our understanding of how BPD features impact social network quality and composition. Our findings suggest BPD features predict greater perceived conflict and criticism, less satisfaction and support, and more overall, romantic, and cut-off relationships in one's network. Larger, less supportive networks containing more partners with whom one has cut off contact and more current or former romantic partners may be especially problematic for individuals with impulsive behavior and difficulties regulating emotions. This is particularly important for those with BPD features, given that positive social relationships may have a protective effect on the experience of negative affect such as anger (Kuhlken et al., 2013) and predict an improved symptom course (Links & Heslegrave, 2000). Acknowledgments The authors thank Star Hess and Victoria Alexander, the undergraduate research assistants who assisted with data collection. References American Psychiatric Association (2013). Diagnostic and statistical manual of mental disorders (5th ed.). Washington, DC: Author. Berenson, K. R., Downey, G., Rafaeli, E., Coifman, K. G., & Paquin, N. L. (2011). The rejection-rage contingency in borderline personality disorder. Journal of Abnormal Psychology, 120, 681–690. http://dx.doi.org/10.1037/a0023335. Bouchard, S., Sabourin, S., Lussier, Y., & Villeneuve, E. (2009). Relationship quality and stability in couples when one partner suffers from borderline personality disorder. Journal of Marital and Family Therapy, 35, 446–455. http://dx.doi.org/10.1111/j. 1752-0606.2009.00151.x. Cheavens, J. S., Lazarus, S. A., & Herr, N. R. (2014). Interpersonal partner choices by individuals with elevated features of borderline personality disorder. Journal of Personality Disorders, 28, 594–606. http://dx.doi.org/10.1521/pedi_2013_27_069. Clifton, A., Pilkonis, P. A., & McCarty, C. (2007). Social networks in borderline personality disorder. Journal of Personality Disorders, 21, 434–441. http://dx.doi.org/10.1521/pedi. 2007.21.4.434. Daley, S., Burge, D., & Hammen, C. (2000). Borderline personality disorder symptoms as predictors of 4-year romantic relationship dysfunction in young women: Addressing issues of specificity. Journal of Abnormal Psychology, 109, 451–460. http://dx.doi.org/ 10.1037//0021-843X.109.3.45. Downey, G., & Feldman, S. I. (1996). Implications of rejection sensitivity for intimate relationships. Journal of Personality and Social Psychology, 70, 1327–1343. http://dx.doi. org/10.1037//0022-3514.70.6.1327. Downey, G., Freitas, A. L., Michaelis, B., & Khouri, H. (1998). The self-fulfilling prophesy in close relationships: Rejection sensitivity and rejection by romantic partners. Journal of Personality and Social Psychology, 75, 545–560. http://dx.doi.org/10.1037/00223514.75.2.545. Gardner, K. J., Qualter, P., Stylianou, M., & Robinson, A. J. (2010). Facial affect recognition in non-clinical adults with borderline personality features: The role of effortful control and rejection sensitivity. Personality and Individual Differences, 49, 799–804. http://dx.doi.org/10.1016/j.paid.2010.07.018. Gunderson, J. G. (2007). Disturbed relationships as a phenotype for borderline personality disorder. The American Journal of Psychiatry, 164, 1637–1640. http://dx.doi.org/10. 1176/appi.ajp.2007.07071125. Hayes, A. F. (2013). Introduction to mediation, moderation, and conditional process analysis: A regression-based approach. New York: The Guilford Press. Hilsenroth, M. J., Menaker, J., Peters, E. J., & Pincus, A. L. (2007). Assessment of borderline pathology using the inventory of interpersonal problems circumplex scales (IIP-C): A comparison of clinical samples. Journal of Clinical Psychology and Psychotherapy, 14, 365–376. http://dx.doi.org/10.1002/cpp.549. Kendler, K. S., Myers, J., & Prescott, C. A. (2005). Sex differences in the relationship between social support and risk for major depression: A longitudinal study of opposite-sex twin pairs. American Journal of Psychiatry, 162, 250–256. http://dx.doi. org/10.1176/appi.ajp.162.2.250. Kuhlken, K., Robertson, C., Benson, J., & Nelson-Gray, R. (2013). The interaction of borderline personality disorder symptoms and relationship satisfaction in predicting affect. Personality Disorders: Theory, Research, and Treatment, 5, 20–25. http://dx.doi.org/10. 1037/per0000013. Links, P. S., & Heslegrave, R. J. (2000). Prospective studies of outcome: Understanding mechanisms of change in patients with borderline personality disorder. The Psychiatric Clinics of North America, 23, 137–150. http://dx.doi.org/10.1016/S0193953X(05)70148-9. Marston, E. G., Hare, A., & Allen, J. P. (2010). Rejection sensitivity in late adolescence: Social and emotional sequelae. Journal of Research on Adolescence, 20, 959–982. http:// dx.doi.org/10.1111/j.1532-7795.2010.00675.x.
Please cite this article as: Lazarus, S.A., et al., Do borderline personality disorder features and rejection sensitivity predict social network outcomes over time?, Personality and Individual Differences (2016), http://dx.doi.org/10.1016/j.paid.2016.02.032
6
S.A. Lazarus et al. / Personality and Individual Differences xxx (2016) xxx–xxx
Miano, A., Fertuck, E. A., Arntz, A., & Stanley, B. (2013). Rejection sensitivity is a mediator between borderline personality disorder features and facial trust appraisal. Journal of Personality Disorders, 27, 442–456. http://dx.doi.org/10.1521/pedi_2013_27_096. Morey, L. C. (1991). Personality assessment inventory: Professional manual. Odessa, FL: Psychological Assessment Resources. Russell, J. J., Moskowitz, D. S., Zuroff, D. C., Sookman, D., & Paris, J. (2007). Stability and variability of affective experience and interpersonal behavior in borderline personality disorder. Journal of Abnormal Psychology, 116, 578–588. http://dx.doi.org/10.1037/ 0021-843X.116.3.578. Selby, E. A., Ward, A. C., & Joiner, T. E. (2010). Dysregulated eating behaviors in borderline personality disorder: Are rejection sensitivity and emotion dysregulation linking mechanisms? International Journal of Eating Disorders, 43, 667–670. http://dx.doi. org/10.1002/eat.20761. Stepp, S. D., Pilkonis, P. A., Yaggi, K. E., Morse, J. Q., & Feske, U. (2009). Interpersonal and emotional experiences of social interactions in borderline personality disorder. The Journal of Nervous and Mental Disease, 197, 484–491. http://dx.doi.org/10.1097/ NMD.0b013e3181aad2e7.
Tragesser, S. L., Lippman, L. G., Trull, T. J., & Barrett, K. C. (2008). Borderline personality disorder features and cognitive, emotional, and predicted behavioral reactions to teasing. Journal of Research in Personality, 42, 1512–1523. http://dx.doi.org/10.1016/ j.jrp.2008.07.003. Trull, T. J., Useda, D., Conforti, K., & Doan, B. (1997). Borderline personality disorder features in nonclinical young adults: 2. Two-year outcome. Journal of Abnormal Psychology, 105, 307–314. http://dx.doi.org/10.1037/0021-843X.106.2.307. Zielinski, M. J., & Veilleux, J. C. (2014). Examining the relation between borderline personality features and social support: The mediating role of rejection sensitivity. Personality and Individual Differences, 70, 235–238. http://dx.doi.org/10.1016/j.paid. 2014.07.005. Zimmerman, M., Chelminski, I., Young, D., Dalrymple, K., & Martinez, J. (2012). Does the presence of one feature of borderline personality disorder have clinical significance? Implications for dimensional ratings of personality disorders. Journal of Clinical Psychiatry, 73, 8–12. http://dx.doi.org/10.4088/JCP.10m06784.
Please cite this article as: Lazarus, S.A., et al., Do borderline personality disorder features and rejection sensitivity predict social network outcomes over time?, Personality and Individual Differences (2016), http://dx.doi.org/10.1016/j.paid.2016.02.032
Contents lists available at ScienceDirect
Personality and Individual Differences journal homepage: www.elsevier.com/locate/paid
Do borderline personality disorder features and rejection sensitivity predict social network outcomes over time? Sophie A. Lazarus ⁎, Matthew W. Southward, Jennifer S. Cheavens Department of Psychology, Ohio State University, United States
a r t i c l e
i n f o
Article history: Received 20 June 2015 Received in revised form 5 November 2015 Accepted 12 February 2016 Available online xxxx Keywords: Borderline personality disorder Social networks Interpersonal functioning Rejection sensitivity Social support
a b s t r a c t Social functioning is routinely understood to be disrupted for those with BPD features; however there is little understanding of how BPD features and BPD-relevant traits impact social network characteristics over time. We hypothesized that BPD features negatively predict social network quality and composition and that rejection sensitivity (RS) would affect these relations. To examine this, a sample of female college students (N = 127) was recruited and followed over one month. BPD features predicted lower ratings of social network quality and aspects of network composition. BPD features exerted an indirect effect through one-week RS on perceived levels of conflict and criticism as well as on number of partners in the network at one-month follow-up. Moderation analyses revealed that BPD features predicted lower social network satisfaction and support at one month for those with high RS, but did not impact satisfaction or support for those with low RS. These results indicate that even non-clinical levels of BPD psychopathology are related to poor social network outcomes. These findings also highlight RS as a potential mechanism by which BPD features predict lower social support and satisfaction and a potential risk factor for higher conflict and criticism within social networks. © 2016 Elsevier Ltd. All rights reserved.
1. Introduction Theoretical and empirical accounts point to difficult interpersonal relationships as a major source of distress for those with borderline personality disorder (BPD; Gunderson, 2007; Hilsenroth, Menaker, Peters, & Pincus, 2007). There is increasing evidence that BPD features are related to both objective (e.g., composition of networks) and subjective (e.g., ratings of network quality) social network constructs. In terms of objective social network outcomes, BPD status is related to smaller social networks, measured by fewer interaction partners per day (Stepp, Pilkonis, Yaggi, Morse, & Feske, 2009), and more BPD features predict the generation of fewer partners available to fulfill the need for social support (Zielinski & Veilleux, 2014). BPD criteria are also related to an increased proportion of conflictual or romantic partners in one's social network (e.g., Clifton, Pilkonis, & McCarty, 2007; Daley, Burge, & Hammen, 2000). Consistent with these compositional differences, BPD features have been associated with less satisfaction and support from romantic partners (Bouchard, Sabourin, Lussier, & Villeneuve, 2009; Daley et al., 2000) and from social partners in general (Zielinski & Veilleux, 2014). However, there is increasing evidence that positive
⁎ Corresponding author at: 181, Psychology Building, 1835 Neil Avenue, Columbus, OH 3210, United States. E-mail address: [email protected] (S.A. Lazarus).
social relationships may result in lower levels of anger, a BPD criterion (Kuhlken, Robertson, Benson, & Nelson-Gray, 2013), and that marriage may positively predict overall global functioning and symptom status over time for those with BPD features (Links & Heslegrave, 2000). Thus, given the protective nature of stable, satisfying relationships and the likelihood that BPD features may interfere with forming or keeping such relationships, it is imperative to determine what predicts high quality relationships at high levels of BPD features. A growing body of research indicates that problematic social network outcomes are associated with BPD features (e.g., Daley et al., 2000; Zielinski & Veilleux, 2014), even in samples not diagnosed with or selected for BPD status. Understanding social network dysfunction across the continuum of BPD features is important given evidence that even minimal levels of BPD pathology have clinical significance (Zimmerman, Chelminski, Young, Dalrymple, & Martinez, 2012) and subclinical BPD features are longitudinally associated with difficulties in academic achievement, mood, and interpersonal functioning (Trull, Useda, Conforti, & Doan, 1997). To date, most existing research has examined the effects of BPD features on social network characteristics without considering the impact of vulnerabilities associated with BPD. Examining specific personality characteristics that may put those with BPD features at increased risk for poor social network outcomes or contribute to specific social network characteristics over time can give us a more precise understanding of the interplay between BPD features and disrupted social networks.
http://dx.doi.org/10.1016/j.paid.2016.02.032 0191-8869/© 2016 Elsevier Ltd. All rights reserved.
Please cite this article as: Lazarus, S.A., et al., Do borderline personality disorder features and rejection sensitivity predict social network outcomes over time?, Personality and Individual Differences (2016), http://dx.doi.org/10.1016/j.paid.2016.02.032
2
S.A. Lazarus et al. / Personality and Individual Differences xxx (2016) xxx–xxx
1.1. Rejection sensitivity (RS) and BPD
1.2. Current study
RS, defined as the tendency to anxiously expect, readily perceive, and react strongly to perceived rejection (Downey & Feldman, 1996), frequently characterizes BPD, yet the two constructs are not synonymous. The empirical association between BPD features and RS has been demonstrated in several studies (e.g., Berenson, Downey, Rafaeli, Coifman, & Paquin, 2011; Tragesser, Lippman, Trull, & Barrett, 2008) and some researchers have investigated RS as a mechanism through which BPD results in problematic outcomes. Among those diagnosed with BPD, Selby, Ward, and Joiner (2010) found that BPD symptoms predicted RS, which predicted dysregulated emotions and eating behavior. In undergraduate samples, RS mediated the relation between BPD features and facial trust appraisal (Miano, Fertuck, Arntz, & Stanley, 2013) and between BPD features and number of social contacts (Zielinski & Veilleux, 2014). However, Zielinski and Veilleux (2014) found that RS did not mediate the relation between BPD features and social support satisfaction. These results can be interpreted to suggest that elevated fears of social rejection in BPD may decrease regulatory ability, leading to impulsive behaviors, less trust of others, and fewer social contacts but that the relation between BPD and satisfaction with one's social support may be a direct effect, not dependent on RS. While the association between BPD and RS across investigations suggests that RS may be important in understanding how and when BPD features predict social network characteristics, these meditational studies were all cross-sectional, limiting our understanding of temporal links. Assessing the composition and quality of social networks can be a daunting task given the number of decisions that must be made to appropriately operationalize and define the construct. One of the first decisions is whether the social network should be examined at one time point or over time. The majority of compositional research to date has examined social network differences between those with BPD and various control participants either at one time point to assess how BPD features and social network qualities are concurrently related (e.g., Clifton et al. (2007)) or over relatively short time periods (e.g., one week; Stepp et al., 2009) to determine how particular interactions impact short-term outcomes such as mood. However, these designs do not permit an investigation of how social networks are affected by BPD features over time. Second, who is considered to be in one's social network is an important aspect of social network research. Some researchers have assessed specific relationships (e.g., romantic partners; Bouchard et al., 2009) while others have examined partners who might be available in particular contexts (e.g., Zielinski & Veilleux, 2014). While these designs are informative, it is also important to assess social networks more broadly, as those with whom the individual frequently interacts may impact their emotional functioning and available support.
We examined how BPD features impacted social network characteristics using ratings of social network relationships over time. Existing research provides some insight into specific characteristics of the social networks of those with BPD and BPD features using various methods, such as daily diary assessment and cross-sectional self-report. However, this research is limited and there is no consensus on the best methods to assess social networks. Given this, we developed a measure, described below, to assess relationships with all partners with whom individuals frequently interact. Based on previous findings, we assessed composition of the social networks (e.g., number of total partners, romantic partners, and partners to whom one had stopped speaking) as well as the quality of each relationship (i.e., satisfaction, support, conflict, and criticism). We tested whether BPD features predicted social network characteristics one month later. We also assessed the influence of RS on these associations. While several studies have examined RS as a mediator of the relations between BPD and socially relevant outcomes (e.g., Miano et al., 2013; Zielinski & Veilleux, 2014), RS also has been tested as a moderator that amplifies risk for deficits in social functioning at higher levels of BPD features (Gardner, Qualter, Stylianou, & Robinson, 2010). We tested RS as both a moderator and a mediator of the relations between BPD features and social network outcomes to clarify whether RS is the mechanism through which these relations exist or a risk factor for poor social outcomes for those with BPD features. Examining interpersonal relationships in this manner allows for a more complete picture of social networks and BPD features and a better understanding of potential risk factors and mechanisms for disrupted network quality. 2. Materials and methods 2.1. Participants and procedure This study was open to all female introductory psychology students. Because BPD is primarily diagnosed in women (American Psychiatric Association, 2013) and there are tend to be differences in social networks (Kendler, Myers, & Prescott, 2005) and RS (Downey, Freitas, Michaelis, & Khouri, 1998) based on gender, we limited participation to female students. At baseline, participants completed consent forms and questionnaires assessing personality variables (i.e., BPD features, RS) and social network characteristics. One week and one month after baseline, participants were emailed a link to complete the same questionnaires. Participants were compensated with either partial course credit or online gift cards (because the semester ended prior to study completion for some participants). The sample consisted of 127 female students at a large Midwestern university. Participants ranged in age from 18 to 32 (M = 19.57,
Table 1 Baseline means, standard deviations, and correlations. Measure a
1. PAI-BOR 2. RSQb 3. Total partners 4. Romantic partners 5. Cut-off partners 6. Satisfaction 7. Support 8. Conflict 9. Criticism
M
SD
21.02 8.86 11.17 1.09 .95 2.42 2.38 1.38 .86
10.07 3.32 6.29 1.94 1.41 .39 .47 .52 .52
1
2
3
4
– .36⁎⁎ .25⁎⁎ −.24⁎⁎ −.23⁎ −.31⁎⁎ −.26⁎⁎
– .23⁎⁎ −.16 −.10 −.09 −.01
5
6
7
8
–
.26⁎⁎ .04 .20⁎ .35⁎⁎ −.42⁎⁎ −.25⁎⁎ .34⁎⁎ .22⁎
– −.14 −.07 .18⁎ −.28⁎⁎ −.16 .40⁎⁎ .29⁎⁎
– −.37⁎⁎ −.33⁎⁎ .08 .01
– −.58⁎⁎ −.21⁎ −.06
– .05 .05
–
.46⁎⁎
⁎ p b .05. ⁎⁎ p b .01. a Personality Assessment Inventory — Borderline subscale. b Rejection Sensitivity Questionnaire.
Please cite this article as: Lazarus, S.A., et al., Do borderline personality disorder features and rejection sensitivity predict social network outcomes over time?, Personality and Individual Differences (2016), http://dx.doi.org/10.1016/j.paid.2016.02.032
S.A. Lazarus et al. / Personality and Individual Differences xxx (2016) xxx–xxx
SD = 2.50) years old and over half were Caucasian (66.1%). The majority of participants were single (66.1%) with a large minority reporting they were in non-cohabitating relationships (27.6%). There were no exclusion criteria. Inclusion criteria were that participants be female, 18 years old or older, and enrolled in an introductory psychology course.
3
Table 2 Baseline PAI-BOR scores predicting one-month SNA outcomes. Poisson regression
df
b
SE
β
χ2
Pseudo R2a
Number of partners Number of romantic partners Number of cut-off partners
1, 125 1, 125 1, 125
.01 .03 .04
b .01 .01 .01
.06 .31 .36
3.71 15.69 12.03
.01 .07 .06
2.2. Study measures 2.2.1. Borderline features The Personality Assessment Inventory — Borderline subscale (PAIBOR; Morey, 1991) is a 24-item, self-report inventory designed to assess BPD features in adults. In this sample, it demonstrated good internal consistency (Cronbach's α = .85) and good test–retest reliability over four weeks (r = .76). 2.2.2. Rejection sensitivity The Rejection Sensitivity Questionnaire (RSQ; Downey & Feldman, 1996) consists of 18 scenarios in which participants imagine requesting something from another person that makes them vulnerable to rejection. Participants rate how anxious they are that the person will respond negatively to the request on a 1 (very unconcerned) to 6 (very concerned) scale, and the likelihood that the person will respond positively to the request on a 1 (very unlikely) to 6 (very likely) scale. Total scores are calculated by taking the inverse of the likelihood ratings, multiplying by the anxiety ratings, and taking the average, so that high scores reflect high anxiety and high likelihood of rejection. In this sample, the scale demonstrated good internal consistency (Cronbach's α = .88) and test–retest reliability over four weeks (r = .81). 2.2.3. Social network Created for this study, the Social Network Assessment (SNA) prompts participants to list people with whom they have interacted two or more times for at least 5 min each time in the previous week. Participants listed up to 24 partners in order of the frequency of interaction. For each partner listed, participants rated their satisfaction, support, conflict, and criticism in the relationship on a 1 (not at all) to 4 (very much) Likert scale. Ratings were averaged for each participant at each time point to produce estimates of global network satisfaction, support, conflict, and criticism. Participants identified whether each partner was a former or current romantic partner (i.e., romantic partner) and if they had “cut off” speaking to each partner in the past month. 2.3. Analytic strategy We first determined the baseline and one-month characteristics of our sample. Poisson regression analyses were conducted to test whether baseline BPD features predicted one-month social network composition count variables (i.e., total number of partners, number of romantic partners, and number of cut-off partners). We conducted OLS linear regression analyses to test whether baseline BPD features predicted onemonth social network quality variables (i.e., satisfaction, support, conflict, and criticism). Mediation analyses were conducted using the PROCESS macro (Hayes, 2013) with 10,000 percentile bootstraps to test whether baseline BPD features exerted an indirect effect on onemonth social network outcomes through one-week RS. Finally, we entered BPD, RS, and their product as predictors of social network outcomes to test whether RS moderated the impact of BPD features on one-month social network characteristics. All analyses were conducted in SPSS version 20. 3. Results 3.1. Baseline characteristics At baseline, participants reported PAI-BOR scores that were consistent with other undergraduate samples (T-score = 48; Morey, 1991)
OLS regression
df
b
SE
β
F
R2
Satisfaction Support Conflict Criticism
1, 125 1, 125 1, 125 1, 125
−.01 −.01 .02 .01
.01 .01 .01 .01
−.26 −.21 .26 .22
8.88 5.58 9.14 6.52
.07 .04 .07 .05
Note. PAI-BOR = Personality Assessment Inventory — Borderline subscale. SNA = Social Network Assessment. Each line represents a separate model. deviance ðPAIBORÞ a Pseudo R2, calculated as 1 deviance . ðintercept onlyÞ
with 6% of the sample scoring above 37 on the PAI-BOR, suggesting clinically relevant BPD features. RSQ scores were also comparable to those found in other undergraduate samples (Downey & Feldman, 1996; Zielinski & Veilleux, 2014). Table 1 includes the means, standard deviations, and correlations among variables at baseline. 3.2. Do baseline BPD features predict one-month social network outcomes? BPD features at baseline predicted interacting with more current or former romantic partners (p b .01), cutting off more partners (p b .01), and marginally predicted more total partners in the network (p = .05) one month later (Table 2). BPD features also predicted more conflict (p b .01) and criticism (p = .01) as well as less satisfaction (p b .01) and support (p = .02) one month later.1,2 3.3. Does RS mediate the impact of BPD on one-month social network outcomes? Next, we attempted to extend Zielinski and Veilleux's (2014) crosssectional findings by testing whether baseline BPD features exerted an indirect effect through one-week RS on one-month social network outcomes. BPD features exerted an indirect effect through RS on the total number of partners (Table 3). BPD features predicted greater RS at one week, which predicted fewer total partners at one month. BPD features did not exert an indirect effect through RS on the number of current or former romantic partners or the number of cut-off partners. Like Zielinski and Veilleux (2014), we found no evidence of mediation for one-month satisfaction or support. However, there was an indirect effect of BPD features through RS on both conflict and criticism at one month. Baseline BPD features predicted greater RS at one week, which predicted greater average perceived conflict and criticism at one month.3 3.4. Does RS moderate the impact of BPD on one-month social network outcomes? Finally, we examined whether RS moderated the impact of BPD features on social network outcomes. The product of baseline BPD features 1 We ran each of the above models controlling for the baseline measure of the respective social network characteristic to test whether BPD features predicted change in each social network outcome. BPD features only predicted one-month change in the number of cut-off partners (χ2(1 N = 128) = 4.92, B = 0.02, p = .03; all other ps N .06). 2 When examined at baseline only, all findings from the Poisson and OLS regression models are interpreted equivalently to the one-month findings (i.e., the signs of the beta weights are in the same direction and result in the same inferential conclusions regarding statistical significance). 3 We ran each of the above models controlling for baseline measures of the respective social network characteristic to test whether BPD features exerted an indirect effect on changes in each social network outcome through RS. BPD features only exerted an indirect effect through RS on changes in criticism (ab = .002, SE = .002, 95% CI: .000, .006).
Please cite this article as: Lazarus, S.A., et al., Do borderline personality disorder features and rejection sensitivity predict social network outcomes over time?, Personality and Individual Differences (2016), http://dx.doi.org/10.1016/j.paid.2016.02.032
4
S.A. Lazarus et al. / Personality and Individual Differences xxx (2016) xxx–xxx
Table 3 One-week RSQ scores mediating baseline PAI-BOR and one-month SNA outcomes. Outcome
ab
SE
LLCI (95%)
ULCI (95%)
Number of partnersa Number of romantic partners Number of cut-off partners Satisfaction Support Conflict Criticism
−.030 b.001 b.001 b −.001 b −.001 .004 .005
.018 .004 .002 .001 .001 .002 .002
−.071 −.008 −.004 −.003 −.003 .001 .001
−.001 .009 .006 .003 .003 .008 .010
Note. The indirect effect (ab) is significantly different from zero if the 95% confidence interval between the lower limit (LLCI) and upper limit (ULCI) does not include zero. RSQ = Rejection Sensitivity Questionnaire. PAI-BOR = Personality Assessment Inventory — Borderline subscale. SNA = Social Network Assessment. ab = −.03, SE = .02, 95% CI: –.060, .003. a Not significantly different from zero when using baseline variables only.
and RS predicted total number of partners at one month (χ2(1 N = 128) = 4.98, b = .002, p = .03). Specifically, BPD features more strongly predicted the total number of partners when RS was high than when RS was low (Fig. 1). However, people low in RS reported more partners than those high in RS, regardless of BPD features. RS did not moderate the impact of BPD features on the number of current or former romantic partners (χ2(1 N = 128) = 3.51, b = .004, p = .06) or cut-off partners (χ2(1 N = 128) = 0.40, b = −.002, p = .53).4 The product of BPD features and RS also predicted one-month social network satisfaction (F(1, 124) = 4.62, b = −.003, p = .03) and support (F(1, 124) = 5.54, b = − .004, p = .02) but not conflict (F(1, 124) = 1.46, b = − .002, p = .23) or criticism (F(1, 124) = 3.67, b = −.003, p = .06). Specifically, BPD features predicted lower social network satisfaction and support at one month for those with high RS, but did not impact satisfaction or support for those with low RS (Fig. 2).5,6
4. Discussion Taken together, these findings help characterize the relations between BPD features and social network outcomes and have implications for understanding the context in which interpersonal difficulties occur for those with heightened features of BPD. First, the results qualify the effect of BPD features on perceived social network quality. Although BPD features negatively predicted both perceived support from and satisfaction with one's social network, this effect was strongest at higher levels of RS. That is, the tendency to anxiously expect and perceive rejection may amplify fears of abandonment, impulsivity, and anger experienced at high levels of BPD features, ultimately eroding perceptions of social support and satisfaction. In a longitudinal study of older adolescents, Marston, Hare, and Allen (2010) found that those high in RS were rated by close friends as less socially accepted and less close one year later. Thus, it is possible that RS impacts social networks from “both ends” by impacting the individual's sense of satisfaction and security in relationships as well as the degree to which others feel close and connected to that individual. Our results suggest that the combination of BPD features and RS may predict perceived deficits in social satisfaction and support as quickly as one month later. 4 Using only baseline variables, the product of BPD features and RS predicted the number of cut-off partners, B = −.006, SE = .003, p = .04. The interpretations of total number of partners and number of romantic partners remained the same using baseline variables. 5 We ran each of the above models controlling for baseline measures of the respective social network characteristic to test whether RS moderated the effect of BPD features on changes in social network outcomes. The interaction of RS and BPD significantly predicted changes in the total number of partners (χ2(1 N = 128) = 23.43, B = .004, p b .01), satisfaction (F(1, 123) = 8.37, B = − .004, p b .01), and support (F(1, 123) = 7.18, B = −.003, p b .01). These effects were in the same direction as the predictive tests. 6 Entering only baseline variables, the product of BPD features and RS did not predict social network satisfaction or support (ps N .30). The interpretations of conflict and criticism remained the same using baseline variables.
Fig. 1. RS moderates the impact of BPD features on the total number of social network partners one month later.
BPD features also positively predicted average conflict and criticism in one's network one month later, and RS mediated this effect. That is, BPD features predicted increased sensitivity to rejection one week later, and RS, in turn, predicted greater perceived conflict and criticism in one's network at the one-month assessment. These findings build on the results of Downey et al. (1998), that high-RS women engage in more negative relationship behaviors (e.g., verbal put-downs, adopting a negative tone of voice, denying responsibility) after conflicts with their partners, leading to relationship erosion. Our findings suggest that women with higher BPD features perceive more conflict and criticism in their network because they are more sensitive to cues of rejection. Future researchers should investigate whether conflict behaviors jointly mediate this relation. Together, these findings suggest that RS may be a mechanism by which BPD features impact perceptions of conflict and criticism as well as a vulnerability that interacts with BPD features leading to perceptions of less satisfaction and support from social networks. Finally, we found that BPD features marginally predicted more social network partners overall and that RS both moderated and mediated this effect in separate models. In the mediation model, BPD features exerted a negative indirect effect on the total number of partners through RS. BPD features predicted greater RS one week later, which predicted fewer partners one month later. In the moderation model, BPD features positively predicted the number of partners at one month when RS was high but were unrelated to the number of partners when RS was low. Notably, those lower in RS tended to report more partners at all levels of BPD features than those higher in RS. These results suggest that among people with the same level of RS, those with more BPD features tend to report more social network partners. It is possible that other aspects of BPD such as impulsivity or novelty seeking, which were not the focus of this study, also contribute to pursuing a greater number of social partners.
Fig. 2. RS moderates the impact of BPD features on average social network satisfaction one month later.
Please cite this article as: Lazarus, S.A., et al., Do borderline personality disorder features and rejection sensitivity predict social network outcomes over time?, Personality and Individual Differences (2016), http://dx.doi.org/10.1016/j.paid.2016.02.032
S.A. Lazarus et al. / Personality and Individual Differences xxx (2016) xxx–xxx
The findings from this study are consistent with previous research showing that BPD features are related to lower relationship satisfaction and more high-conflict relationships (Clifton et al., 2007; Zielinski & Veilleux, 2014). While previous researchers have found evidence of decreased satisfaction in several specific interpersonal relationships (Bouchard et al., 2009; Stepp et al., 2009; Zielinski & Veilleux, 2014), a significant strength of the current study is the assessment of a wide, unselected range of partners over time. Assessing social network partners in this way may also explain why we found that BPD features were positively related to the total number of partners while other researchers found that BPD was related to fewer supportive partners (Zielinski & Veilleux, 2014) and fewer social interactions (Stepp et al., 2009). BPD features may actually be related to interacting with more partners, but perhaps also to the perception of those partners as less supportive or reliable for daily needs. This is consistent with research showing that BPD features are related to a preference for novel over familiar partners in a behavioral laboratory task (Cheavens, Lazarus, & Herr, 2014). The presence of more interaction partners in the network, or the behavior of seeking out new partners, may interfere with the ability to invest in individual relationships. It is important to consider details such as specific methodology for collecting network information when considering different findings regarding social network qualities. These results also establish the relation between BPD features and perceived conflict and criticism in an undergraduate sample. These findings build on those of Russell, Moskowitz, Zuroff, Sookman, and Paris (2007) who found that individuals with BPD reported engaging in more quarrelsome behaviors over 20 days than healthy controls. The significance of this finding in a non-clinical sample suggests that even elevated features of BPD among those with normal variation in symptoms of the disorder are related to networks with higher levels of conflict and criticism and reports of quarrelsome behavior are consistent with reports of conflictual relationships. Higher BPD features were also associated with more network members who were current or former romantic partners and more partners who were cut off in the past month. The presence of more conflictual partners may contribute to greater perceived conflict and criticism among network partners. However, while RS mediated the relation between BPD features and network conflict and criticism, it did not mediate or moderate the relations between BPD features and romantic or cut-off partners. Anxiously expecting and perceiving rejection may more strongly influence perceptions of partners than the type of partners in a person's network. Some aspects of the current study limit the strength of the conclusions we can draw and point to areas for future research. First, while it is important to establish the relations between BPD features, RS, and various social network outcomes in those with normal variation in these features, these relations may be different for individuals diagnosed with BPD. Perhaps in a clinical sample, where the relation between BPD and RS is likely stronger, our ability to detect the role of RS on social networks may be limited. However, the relation between BPD and RS found in clinical samples (e.g., Selby et al., 2010) suggests that these constructs are far from redundant, and this should be examined empirically. In addition, RS is likely not the only personality characteristic that increases the risk of poor social network outcomes for those with BPD features. Affective instability, interpersonal aggression, or emotion regulation difficulties may also interact with BPD diagnosis or features to impact the quality or composition of social networks. Examining other potential vulnerabilities in those with BPD features would contribute to our understanding of the processes by which network differences develop or are maintained. In addition, although information on specific social network partners was obtained at baseline and one month, we cannot assume that the same partners define each participant's network at each time point. Thus, while we report the average social network quality and composition for each participant, we can only say that change over time is due to global social network changes, not changes in specific relationships. In future studies, it would be useful to assess the same partners at each
5
assessment to evaluate change within a given relationship as well as on average in the network. Finally, while the longitudinal design is a strength of this study, longer follow-up periods may reveal more variance in social network outcomes and increase the opportunity to determine how changes in social networks are related to BPD features. This study contributes to our understanding of how BPD features impact social network quality and composition. Our findings suggest BPD features predict greater perceived conflict and criticism, less satisfaction and support, and more overall, romantic, and cut-off relationships in one's network. Larger, less supportive networks containing more partners with whom one has cut off contact and more current or former romantic partners may be especially problematic for individuals with impulsive behavior and difficulties regulating emotions. This is particularly important for those with BPD features, given that positive social relationships may have a protective effect on the experience of negative affect such as anger (Kuhlken et al., 2013) and predict an improved symptom course (Links & Heslegrave, 2000). Acknowledgments The authors thank Star Hess and Victoria Alexander, the undergraduate research assistants who assisted with data collection. References American Psychiatric Association (2013). Diagnostic and statistical manual of mental disorders (5th ed.). Washington, DC: Author. Berenson, K. R., Downey, G., Rafaeli, E., Coifman, K. G., & Paquin, N. L. (2011). The rejection-rage contingency in borderline personality disorder. Journal of Abnormal Psychology, 120, 681–690. http://dx.doi.org/10.1037/a0023335. Bouchard, S., Sabourin, S., Lussier, Y., & Villeneuve, E. (2009). Relationship quality and stability in couples when one partner suffers from borderline personality disorder. Journal of Marital and Family Therapy, 35, 446–455. http://dx.doi.org/10.1111/j. 1752-0606.2009.00151.x. Cheavens, J. S., Lazarus, S. A., & Herr, N. R. (2014). Interpersonal partner choices by individuals with elevated features of borderline personality disorder. Journal of Personality Disorders, 28, 594–606. http://dx.doi.org/10.1521/pedi_2013_27_069. Clifton, A., Pilkonis, P. A., & McCarty, C. (2007). Social networks in borderline personality disorder. Journal of Personality Disorders, 21, 434–441. http://dx.doi.org/10.1521/pedi. 2007.21.4.434. Daley, S., Burge, D., & Hammen, C. (2000). Borderline personality disorder symptoms as predictors of 4-year romantic relationship dysfunction in young women: Addressing issues of specificity. Journal of Abnormal Psychology, 109, 451–460. http://dx.doi.org/ 10.1037//0021-843X.109.3.45. Downey, G., & Feldman, S. I. (1996). Implications of rejection sensitivity for intimate relationships. Journal of Personality and Social Psychology, 70, 1327–1343. http://dx.doi. org/10.1037//0022-3514.70.6.1327. Downey, G., Freitas, A. L., Michaelis, B., & Khouri, H. (1998). The self-fulfilling prophesy in close relationships: Rejection sensitivity and rejection by romantic partners. Journal of Personality and Social Psychology, 75, 545–560. http://dx.doi.org/10.1037/00223514.75.2.545. Gardner, K. J., Qualter, P., Stylianou, M., & Robinson, A. J. (2010). Facial affect recognition in non-clinical adults with borderline personality features: The role of effortful control and rejection sensitivity. Personality and Individual Differences, 49, 799–804. http://dx.doi.org/10.1016/j.paid.2010.07.018. Gunderson, J. G. (2007). Disturbed relationships as a phenotype for borderline personality disorder. The American Journal of Psychiatry, 164, 1637–1640. http://dx.doi.org/10. 1176/appi.ajp.2007.07071125. Hayes, A. F. (2013). Introduction to mediation, moderation, and conditional process analysis: A regression-based approach. New York: The Guilford Press. Hilsenroth, M. J., Menaker, J., Peters, E. J., & Pincus, A. L. (2007). Assessment of borderline pathology using the inventory of interpersonal problems circumplex scales (IIP-C): A comparison of clinical samples. Journal of Clinical Psychology and Psychotherapy, 14, 365–376. http://dx.doi.org/10.1002/cpp.549. Kendler, K. S., Myers, J., & Prescott, C. A. (2005). Sex differences in the relationship between social support and risk for major depression: A longitudinal study of opposite-sex twin pairs. American Journal of Psychiatry, 162, 250–256. http://dx.doi. org/10.1176/appi.ajp.162.2.250. Kuhlken, K., Robertson, C., Benson, J., & Nelson-Gray, R. (2013). The interaction of borderline personality disorder symptoms and relationship satisfaction in predicting affect. Personality Disorders: Theory, Research, and Treatment, 5, 20–25. http://dx.doi.org/10. 1037/per0000013. Links, P. S., & Heslegrave, R. J. (2000). Prospective studies of outcome: Understanding mechanisms of change in patients with borderline personality disorder. The Psychiatric Clinics of North America, 23, 137–150. http://dx.doi.org/10.1016/S0193953X(05)70148-9. Marston, E. G., Hare, A., & Allen, J. P. (2010). Rejection sensitivity in late adolescence: Social and emotional sequelae. Journal of Research on Adolescence, 20, 959–982. http:// dx.doi.org/10.1111/j.1532-7795.2010.00675.x.
Please cite this article as: Lazarus, S.A., et al., Do borderline personality disorder features and rejection sensitivity predict social network outcomes over time?, Personality and Individual Differences (2016), http://dx.doi.org/10.1016/j.paid.2016.02.032
6
S.A. Lazarus et al. / Personality and Individual Differences xxx (2016) xxx–xxx
Miano, A., Fertuck, E. A., Arntz, A., & Stanley, B. (2013). Rejection sensitivity is a mediator between borderline personality disorder features and facial trust appraisal. Journal of Personality Disorders, 27, 442–456. http://dx.doi.org/10.1521/pedi_2013_27_096. Morey, L. C. (1991). Personality assessment inventory: Professional manual. Odessa, FL: Psychological Assessment Resources. Russell, J. J., Moskowitz, D. S., Zuroff, D. C., Sookman, D., & Paris, J. (2007). Stability and variability of affective experience and interpersonal behavior in borderline personality disorder. Journal of Abnormal Psychology, 116, 578–588. http://dx.doi.org/10.1037/ 0021-843X.116.3.578. Selby, E. A., Ward, A. C., & Joiner, T. E. (2010). Dysregulated eating behaviors in borderline personality disorder: Are rejection sensitivity and emotion dysregulation linking mechanisms? International Journal of Eating Disorders, 43, 667–670. http://dx.doi. org/10.1002/eat.20761. Stepp, S. D., Pilkonis, P. A., Yaggi, K. E., Morse, J. Q., & Feske, U. (2009). Interpersonal and emotional experiences of social interactions in borderline personality disorder. The Journal of Nervous and Mental Disease, 197, 484–491. http://dx.doi.org/10.1097/ NMD.0b013e3181aad2e7.
Tragesser, S. L., Lippman, L. G., Trull, T. J., & Barrett, K. C. (2008). Borderline personality disorder features and cognitive, emotional, and predicted behavioral reactions to teasing. Journal of Research in Personality, 42, 1512–1523. http://dx.doi.org/10.1016/ j.jrp.2008.07.003. Trull, T. J., Useda, D., Conforti, K., & Doan, B. (1997). Borderline personality disorder features in nonclinical young adults: 2. Two-year outcome. Journal of Abnormal Psychology, 105, 307–314. http://dx.doi.org/10.1037/0021-843X.106.2.307. Zielinski, M. J., & Veilleux, J. C. (2014). Examining the relation between borderline personality features and social support: The mediating role of rejection sensitivity. Personality and Individual Differences, 70, 235–238. http://dx.doi.org/10.1016/j.paid. 2014.07.005. Zimmerman, M., Chelminski, I., Young, D., Dalrymple, K., & Martinez, J. (2012). Does the presence of one feature of borderline personality disorder have clinical significance? Implications for dimensional ratings of personality disorders. Journal of Clinical Psychiatry, 73, 8–12. http://dx.doi.org/10.4088/JCP.10m06784.
Please cite this article as: Lazarus, S.A., et al., Do borderline personality disorder features and rejection sensitivity predict social network outcomes over time?, Personality and Individual Differences (2016), http://dx.doi.org/10.1016/j.paid.2016.02.032