| This study evaluated the internal and external validity of an adolescent measure of psychopathy, the Psychopathy Checklist: Youth Version (PCL-YV). The sample consisted of 227 adolescents residing in correctional or psychiatric inpatient settings. Internal validity was assessed through a series of confirmatory factor analyses comparing five different models of psychopathy that have been proposed in the literature: (1) Original two-factor model (Harpur et al., 1989); (2) Cooke and Michie (2001) three-factor model; (3) Hare (2003) four-factor model; (4) PCL:YV manual (Forth et al, 2003) three-factor model, (5) PCL:YV manual four-factor model. Each factor model was evaluated using individual items, as well as items composites (also known as parcels or testlets) as indicators. Multiple fit indices supported only Cooke and Michie's three-factor model (chi2 = 4.83, df = 6, CFI = 1.00, GFI = .99, AGFI = .98, RMSEA = .000) and Hare's (2003) four-factor model (chi 2 = 30.64, df = 14, CFI = .96, GFI = .97, AGFI = .91, RMSEA = .073). Model fit improved substantially with the use of a parceling approach. To evaluate the external validity of the competing factor models, each factor was correlated with an independent, self-report measure of personality and psychopathology, the Millon Adolescent Clinical Inventory (MACI). There was a consistent and theoretically meaningful pattern of correlations between each PCL:YV factor and MACI factors and scales, with statistically significant correlations ranging from .14 to .46. Of special interest, Hare's fourth factor, derived from the antisocial items omitted in the Cooke and Michie model, was significantly correlated with three of seven MACI factors (Antisocial-Delinquent, Aggressive, and Interpersonal Problems). Stepwise regression analyses indicated that the Antisocial factor provided incremental validity beyond the other three psychopathy factors in predicting scores on the MACI Antisocial-Delinquent factor. |
PDF Full Text Request