A confirmatory and exploratory investigation of the information processing and cognitive mediation clusters of the Rorschach Comprehensive System

This study examined Rorschach variables comprising the Information Processing IP and the Cognitive Mediation (CM) Clusters of the Exner Comprehensive System within a large inpatient adolescent population. The IP and CM clusters are hypothesized by Exner to provide data regarding how an individual takes in and translates stimulus information. Variables of interest were dichotomized into categorical subgroups in accordance with Exner's hypotheses and examined in relation to available personality, cognitive, and symptom measures. The analyses conducted permitted general cross-validation of predicted interpretive postulates, and facilitated exploratory investigations of unanticipated relationships. The sample consisted of 998 adolescent psychiatric inpatients (age 13--17 years) referred for comprehensive psychodiagnostic and neuropsychological testing. Adolescents that provided a valid Rorschach under the guidelines set forth by Exner (1993), and completed the majority of the test battery were included in the study. Rorschach variables of interest were dichotomized using Exner's suggested cut scores and compared to other variables in the test battery. Percent agreement for dichotomized variables ranged from 90% to 100%. Kappas ranged from .53 to 1.00. 156 one-way confirmatory ANOVAS were conducted between dichotomous Rorschach variables and other cognitive and psychological variables. There were 48 (31%) significant findings (p < .05) consistent with hypotheses derived from Exner. The variables that performed as predicted were: Lambda, The Hypervigilance Index (HVI), Developmental Quality (DQ), and Confabulation (CFB). Partial support was also found for Perseveration (PSV). Exploratory Pearson correlations and chi-squares yielded significant (p < .01) unanticipated information about IP and CM variables. Specifically, High Lambda (L > .99) and underincorporation (Zd < --3.0) were negatively associated with self-report measures of psychopathology, and positively associated with clinician ratings of behavioral difficulties. Low Lambda (L < .31), overincorporation (Zd > +.30), and a positive HVI index were positively associated with self-report measures of psychiatric symptoms and clinician ratings of internalizing disorders. This study represents an initial exploration into the meaning of IP and CM variables with regard to adolescents. Future investigations should utilize optimal cut scores and naturally occurring distributions to subdivide samples, and focus on normal controls and outpatient adolescents in an attempt to cross-validate the relationships found in this study.