Predictive Validity of Adolescent Callous-Unemotional Traits and Conduct Problems with Respect to Adult Outcomes: High- and Low-Risk Samples

In the present study, we used mean scores on the Antisocial Process Screening Device [64] to be consistent with other studies using these samples. However, most other research using this measure reports sum scores. When we convert our mean scores into sum scores, we find the following for each of our samples: FT: M = 3.74, SD = 2.20; CDP: M = 2.99, SD = 1.75, which are consistent with other studies. For example, in a sample of community adolescents (M = 13.40 years), Muñoz and Frick [85] found similar sum scores to the CDP sample and lower than the FT sample across three time points approximately 1 year apart (M = 2.41–2.70; SD = 1.94–2.05). Comparably, in another sample of community adolescents (aged 12–14 years), Shaffer et al. [86] found similar self-reported sum scores (M = 2.95, SD = 1.74).

Although CDP also has parent-report information for conduct problems, we elected to use self-report to align with our measure of CU traits (i.e., CDP only has self-report information for CU traits).

Items with content overlap include Antisocial Process Screening Device [65] ‘feel bad when do something wrong’, reverse coded, for CU traits, and CBCL [62] ‘doesn’t feel guilty about misbehavior’ for conduct problems.

We also repeated analyses with Adult Self-Report [68] continuous aggressive behavior, delinquency/rule-breaking, attention problems, anxious/depressed, withdrawn, and somatic complaints narrowband scales (see Supplementary Table S7). Without Bonferroni correction, for FT, CU traits did not predict any outcome and conduct problems predicted every outcome. For CDP, CU traits predicted aggressive behavior, attention problems, anxious/depressed, and withdrawn scores. Conduct problems predicted every outcome.

We also repeated analyses with just the final time points, i.e., age 32 for FT and 34 for CDP. Without Bonferroni correction, many significant effects remained the same. For FT, CU traits negatively predicted high school graduation (OR = 0.43). There were no other significant effects for CU traits. Conduct problems positively predicted externalizing (β = 0.22) and internalizing (β = 0.14) psychopathology, as well as endorsement of clinical levels of ASPD (OR = 1.38), depression (OR = 1.43), and avoidant personality (OR = 1.48); and negatively predicted the general health index (β = − 0.13), happiness (β = − 0.12), high school graduation (OR = 0.74), and college completion (OR = 0.45). For CDP, CU traits positively predicted internalizing psychopathology (β = 0.16) and endorsement of clinical levels of depression (OR = 5.80), and negatively predicted happiness (β = − 0.13), strength (β = − 0.12) and college completion (OR = 0.23). Conduct problems positively predicted externalizing (β = 0.42) and internalizing (β = 0.21) psychopathology; endorsement of clinical levels of ASPD (OR = 2.40), ADHD (OR = 3.18), anxiety (OR = 1.93), depression (OR = 1.79), avoidant personality (OR = 2.02), binge drinking (OR = 2.04), and partner violence (β = 0.17). Conduct problems negatively predicted the general health index (β = − 0.13) and happiness (β = − 0.20).