The present study evaluated the self-report version of the Inventory of Callous-Unemotional Traits (ICU-SR) in terms of reliability, concurrent validity, and correlation with salivary oxytocin levels, a potential biomarker of CU traits. 67 socially at-risk male adolescents (mean 16.2 years) completed the ICU-SR, ICU teacher-version (ICU-TR), Strengths and Difficulties Questionnaire, and their medical files were coded for previous antisocial acts using Brown-Goodwin Lifetime Aggression Scale. Salivary samples were assayed for oxytocin. The reliability of ICU-SR was lower (alpha = 0.71) than ICU-TR (alpha = 0.86). ICU-SR mean score was significantly lower than ICU-TR (M = 25.29, SD = 8.02; M = 33.14, SD = 9.47). ICU-TR but not ICU-SR, significantly correlated with history of antisocial acts (r = 0.40). Two-way analysis of variance showed a significant effect of conduct disorder and oxytocin on ICU-TR but not ICU-SR [F(1,59) = 6.53; F(1,59) = 6.08], and a significant interaction only for ICU-TR [F(1,59) = 2.89]. Subjective self-reports of CU traits may be less reliable and valid than teachers’ reports.