Teachers in a private special education school for students with learning and/or adjustment difficulties completed Bristol Social Adjustment Guides (BSAGs), an observation scale for identifying maladaptive classroom behaviors, for 157 students (7-21 years old). Rorschachs were administered to the same group of students. Data from each test were reduced via factor analyses, and factor structures were chosen which demonstrated both statistical consistency and psychological integrity. Standardized factor scores were subjected to a series of canonical correlations to determine degree of common variance. Zero-order correlations were computed among relevant BSAG, Rorschach, and background variables. Finally, a series of multiple-discriminant analyses were conducted comparing hit-rates for correctly classifying students by special education category and psychiatric diagnosis. No significant canonical correlations were found between Rorschach and BSAG factors. However, both sets of factors extracted appeared unique for this sample. Inclusion and exclusion of certain Rorschach variables were consistent with current research. Discriminant analyses revealed that the BSAG and Rorschach had comparable hit rates (53%) in correctly classifying students by special education category. While Rorschach contents and determinants were most accurate when categorizing students in terms of psychiatric diagnosis, potential use of Rorschach and observation data in the derivation of all of these categories could have confounded findings. (Author/CL)