The aims of the current study were to: 1) identify patterns of agreement between parent-adolescent dyads on reports of adolescent borderline personality features utilizing latent class analysis (LCA) and 2) examine the clinical implications of class membership for indices of psychiatric severity and internal psychological resources. The sample included 643 adolescent inpatients. Borderline personality features were assessed by both adolescents and parents using the Borderline Personality Features Scale - Child (BPFS-C; Crick et al. 2005) and Borderline Personality Features Scale - Parent (BPFS-P; Sharp et al. 2010), respectively. Following recommended statistical approaches for evaluating rater concordance, LCA was utilized to identify distinct classes of parent-adolescent dyads based on concordance/discrepancy in BPFS reports. The subsequent classes were then related to outcome measures of psychiatric severity and internal psychological resources (emotion regulation and experiential acceptance). LCA identified 3 classes of parent-adolescent dyads: 2 convergent classes demonstrating BPFS-P and BPFS-C agreement at a moderate and high level and a divergent class consisting of dyads reporting clinically significant scores on the BPFS-P but clinically negligible BPFS-C scores. Both convergent classes evidenced higher rates of psychiatric severity and lower access to internal resources. The current study is the first to use LCA to examine the relation between informant concordance on reports of DSM-based adolescent borderline pathology in a clinical sample. The significance of the discrepancies within and between classes is discussed with relation to psychosocial outcomes, the diagnosis of borderline personality disorder and implications for what it means when parents and adolescents disagree.