FABBS Undergraduate Research Excellence Award

Carly Miron, University of Pennsylvania, Department of Psychology

“Longitudinal Association between Callous-Unemotional Traits and Friendship Quality among Adjudicated Adolescents”

Carly Miron, University of Pennsylvania, Department of Psychology


Youth with callous-unemotional (CU) traits show severe antisocial behavior and deficits in socioaffiliative processes, such as empathy and prosocial behavior. Adolescence represents a critical developmental window when these socioaffiliative processes support the development of friendships. However, few studies have explored the relationship between CU traits and adolescent friendship quality. In the current study, we used data from the Pathways to Desistance dataset to examine associations between CU traits and friendship quality at three assessment points each separated by 6 months. The sample included adolescents who had interacted with the justice system (baseline age, M=16.04, SD=1.14; N=1354; 13.6% female). CU traits were assessed using the callousness scale of the self-reported Youth Psychopathic Traits Inventory and friendship quality using the Quality of Relationships Inventory. Models accounted for aggression, impulsive-irresponsible traits, grandiose-manipulative traits, age, gender, and race. At every assessment, CU traits were uniquely related to lower friendship quality. Moreover, across the first two assessments, CU traits were related to decreases in friendship quality, while lower friendship quality simultaneously predicted increases in CU traits across the same period. Interventions for CU traits could include specific modules that target socioaffiliative processes and promote capacity for friendship-building, including empathic listening and other-orientated thinking.


A large body of research has established the importance of adolescent friendships for the development of successful adult relationships and socialization. In particular, friendships provide a framework for adolescents to learn and refine socioemotional skills that support the development of positive relationships in adulthood and help them to cement cooperation skills and an ability to understand and empathize with the perspectives of others. Two core interpersonal and affiliative processes that contribute to the development of healthy, positive, and high-quality friendships during adolescence are empathy (i.e., sharing in the perspectives and feelings of others) and guilt (i.e., aversive moral emotion that accompany knowledge that one has committed a misdeed that could harm a relationship or friendship). There are likely bidirectional relationships between empathy and guilt and the development and maintenance of friendships during adolescence. That is, showing caring and empathic behaviors might help in promoting or maintain social relationships during adolescence. Equally, having strong existing friendships might help to increase prosocial and caring behaviors because adolescents learn or mimic such skills from friends who model these types of positive social behaviors or who positively reinforce, validate, or expect such behavior.

Importantly, youth with callous-unemotional (CU) traits show striking deficits in empathy, guilt, and prosocial behaviors. CU traits are also related to severe and chronic forms of antisocial behavior and psychopathic traits, including interpersonally harmful forms of antisocial behavior, such as violence, aggression, and bullying. Surprisingly few studies, however, have examined the relationship between friendship quality and CU traits during adolescence, which could provide further insight into the social factors that could contribute to the development and maintenance of CU traits or that could inform novel social intervention targets to help reduce risk for CU traits and antisocial behavior across adolescence. Further, while plausible that children high on CU traits are at risk for poorer friendship quality over time, it is also possible that experiencing poorer friendship quality could contribute to chronicity in CU traits over time. Within this framework, and as outlined above, an absence of supportive friendships could mean that adolescents have fewer opportunities to enact empathic, prosocial, and caring behaviors within the context of social relationships, further increasing the risk of peer rejection and poor friendship quality, ultimately compounding risk for CU traits in a bidirectional cascade. To explore this potential reciprocity in links between CU traits and poor friendship quality during adolescence, the current study addressed these questions by examining cross-lagged relationships between CU traits and friendship quality over three assessment periods during adolescence.  


We used data from the Pathways to Desistance project, a multisite, longitudinal study of 1354 juvenile offenders. Participants in the current study were youths who had been charged with a felony or serious non-felony offense (e.g., misdemeanor weapons offense or sexual assault) at a court appearance in Philadelphia, PA (N=700) or Phoenix, AZ (N=654). Baseline interviews were conducted an average of 36.9 days (SD=20.6) after court appearances. Interviewers and participants sat side-by-side facing a computer, and questions were read aloud to avoid problems caused by reading difficulties, with answers also being given aloud. Participants completed follow-up interviews at 6, 12, 18, 24, 30, 36, 48, 60, 72 and 84 months past baseline. The current study included data from only the first three follow-up interviews to retain an adolescent population. In the current study, “Time 1”, “Time 2”, and “Time 3” refer to the interviews that took place 6 months, 12 months, and 18 months post-baseline, respectively.

The current study used adolescent reports of friendship quality assessed using the 10-item Friendship Quality Scale. Youth also completed the Youth Psychopathic Traits Inventory (YPI), a 50-item self-report measure of psychopathy that forms three dimensions: grandiose-manipulative (e.g. dishonest charm, grandiosity, lying, manipulation), impulsive-irresponsible (e.g. impulsiveness, irresponsibility, thrill-seeking), and CU traits (e.g. remorselessness, callousness, unemotionality). Note that we included the impulsive-irresponsible and grandiose-manipulative dimensions in analyses to ensure specificity of any effects of CU traits. Similarly, adolescents rated their aggression to establish specificity of any effects of CU traits on friendship quality, over and above the influence of aggression. Youth completed the suppression of aggression subscale of the Weinberger Adjustment Inventory (WAI), a widely-used assessment of an individual’s social-emotional adjustment within the context of external constraints.

Finally, we modeled the effects of relevant demographic covariates, including gender, race, and age. In addition, to account for the fact that interviews with participants varied in location, with some carried out in participants’ homes and others in a detention center or other locked facility, which could impact the ability to form meaningful interpersonal connections and friendships with peers, we created a binary location variable, which we accounted for in relevant models (0=at home/somewhere else in the community vs. 1=in a detention center, jail, or other locked facility). The proportion of the sample in a detention center, jail, or other locked facility each assessment wave was as follows: time 1=50%, time 2=37%, time 3=32%).

While the amount of missing data was low (covariance coverage =.71-.96), we used full information maximum likelihood (FIML) to accommodate any missing data and provide less biased estimates than likewise or pairwise deletion. First, within multivariate regression analyses, we explored the associations between CU traits and friendship quality in adolescence at each time point, controlling for gender, age, race, location of interview, grandiose-manipulative traits, impulsive-irresponsible traits, and aggression. Second, to examine the bidirectional relationship between CU traits and friendship over time, we explored a longitudinal, cross-lagged panel model with gender, race, and age included as covariates.


Within all three cross-sectional regression models, we found that CU traits were significantly related to lower friendship quality at Time 1 (β =-.11, p<.01), Time 2 (β=-.25, p<.01), and Time 3 (β =-.11, p<.01), controlling for gender, age, race, grandiose-manipulative and impulsive-irresponsible traits, location, and aggression at each time point. Within the longitudinal models, and after accounting for stability in friendship quality and CU traits (range, β=.48-.60, p<.001), we found evidence for reciprocal effects. Specifically, lower friendship quality at Time 1 predicted higher CU traits at Time 2 (β=-.07, p<.05), while simultaneously, higher CU traits at Time 1 predicted lower friendship quality at Time 2 (β =-.07, p<.05). Lower friendship quality at Time 2 also predicted higher CU traits at Time 3 (β =-.07, p< .05), although higher CU traits at Time 2 was unrelated to lower friendship quality at Time 3.


We provide consistent evidence that higher levels of CU traits are uniquely related to poorer friendship quality at three different assessments points in adolescence separated by 6 months each. Moreover, the relationship between CU traits and lower friendship quality remained significant after accounting for grandiose-manipulative and impulsive-irresponsible traits, aggression, setting, and other demographic factors, suggesting that something specific about the characteristic features of CU traits, as opposed to psychopathic traits or aggression more broadly, impairs the development of healthy friendships and social affiliation with others. In particular, our results speak to the importance of empathy, prosociality, guilt, and cooperation, essential socioemotional skills that are impaired among youth with CU traits, to the maintenance of high-quality friendships. We also found preliminary evidence for a bidirectional relationship between CU traits and friendship quality over time, but only for the first two assessment points. In particular, across the first two assessment points, poorer friendship quality predicted increases in CU traits over time while CU traits simultaneously predicted decreases in friendship quality. Our results have important translational implications. Specifically, future intervention and treatment efforts involving youth with CU traits could provide the opportunity for adolescents to participate in supportive friendship activities, where they can improve their ability to prioritize the feelings of others, harness empathic listening skills, and practice conflict resolution. These efforts are sorely needed in light of robust evidence linking CU traits to the most serious and chronic forms of youth violence and antisocial behavior. By targeting socioaffiliative mechanisms that support friendship building, as suggested in this study, we can help to promote social engagement with others, empathy, and other-orientated motivation among high-CU youth, which can ultimately reduce risk for antisocial, violent, or criminal behavior.

Impact Statement

Youth with callous-unemotional (CU) traits (i.e., low empathy, guilt, and prosocial behavior) are at heightened risk for developing severe and chronic forms of antisocial behavior and violence, which together constitute a major and costly public health problem. This study showed that CU traits were consistently and reciprocally related to lower friendship quality, even after accounting for co-occurring aggression, impulsiveness, and manipulativeness. The findings suggest that we can help to reduce CU traits, and by extension mitigate the development of youth violence and antisocial behavior, by developing interventions that promote capacity for friendship-building in youth, including by teaching empathic listening and other-orientated thinking skills.