Person-oriented methods in the diagnosis of child behavior
Hendry, Cheryl Dianne Nemeth
MetadataShow full item record
Dimensional, person-oriented classification methods have preliminary evidence of utility for the classification of child psychopathology. Using cluster analysis on scores from the Behavior Assessment System for Children (BASC; Reynolds & Kamphaus, 1992) Teacher Rating Scales for Children (TRS-C) for a normative sample, Huberty, DiStefano, and Kamphaus (1997) developed a seven-cluster typology of child behavior in schools. They found these clusters to be internally valid; therefore, Kamphaus, Huberty, DiStefano, and Petoskey (1997) afforded them substantive interpretation. Similarly, Kamphaus, Petoskey, Cody, Rowe, Huberty, and Reynolds (1999) cluster analyzed scores on the BASC Parent Scales for Children (PRS-C) for a normative sample, resulting in a nine-cluster typology of child behavior. However, in order for dimensional classification systems such as these to be useful in psychological science, they must be able to reliably differentiate child characteristics and illustrate degrees of functional impairment. The purpose of this study was to externally validate the BASC TRS-C typology using a sample of 200 clinic-referred children. BASC TRS-C scores for this independent sample were cluster analyzed and cross-validated. A six-cluster typology was selected for further analysis and interpretation. Each cluster was characterized according to degree of functional impairment as represented by external indicators of behavioral adjustment, school performance, and cognitive development. Additionally, risk factors and outcomes for each cluster were identified. A cross-classification analysis using the BASC TRS-C normative solution was also conducted. Suggestions for further research involving larger sample sizes consisting of varying demographic profiles are offered.