Date of Award


Degree Name

Doctor of Philosophy



First Advisor

Dr. Chris Koronakos

Second Advisor

Dr. Byron Rourke

Third Advisor

Dr. Frederick Gault

Fourth Advisor

Dr. Thomas Van Den Abell


The purpose of this exploratory and confirmatory contrasted-group classification study was to determine if specifically predicted (Hypothesis 1) and externally valid (Hypotheses 3 and 4) subtypes of childhood externalizing disorders could be identified, using previously obtained neuropsychological (NP), socio-emotional (SE), and historical variables. Cross-classification validity was evaluated to determine if subjects NP and SE subtype memberships could be predicted using either their SE and NP subtype memberships or their SE and NP variable scores, respectively (Hypothesis 2). Subjects (n = 117) were 6-to-14-year-old boys and girls with normal intelligence, elevated scores on the Delinquency and/or Hyperactivity scales of the Personality Inventory for Children (Wirt, Lachar, Klinedinst, & Seat, 1977, 1984), but without a post-natal history of brain damage.

As predicted in Hypothesis 1, the existence of several of the SE and NP subtypes was supported. Contrary to Hypothesis 2, moderate cross-classification validity was obtained for accurately classifying subjects to their NP and SE subtypes using their SE or NP variable scores, but not their SE or NP subtype memberships, respectively. With regard to Hypotheses 3 and 4, limited external validity for the NP and SE subtypes was found when considering their multivariate mean differences across SE, NP, psychosocial, developmental, academic, and demographic variables. Contrary to Hypotheses 3 and 4, few univariate mean differences were obtained across these variables, although some significant univariate pair-wise mean differences confirmed research predictions. Consistent with Hypotheses 3 and 4, some nonsignificant contingency table results were obtained across these NP or SE subtypes across various medical, psychosocial, and demographic variables. Using other medical, psychosocial, and educational variables, no significant differences in the rates of these variables were obtained across NP or SE subtypes which disconfirmed other predictions. These results were compared to previous studies of childhood externalizing disorders. The methodological value of this study and current cluster analytic research were reviewed.

Access Setting

Dissertation-Open Access