Method factors associated with assessment of child behavior: fact or artifact?

A sample of 64 phobic children between the ages of 6 and 16, with their general population matched controls, were rated by parents on the Louisville Behavior Check List and Louisville Fear survey and by School Behavior check list. The analysis of ratings by factor analysis revealed three instrument factors while a multiple discriminant analysis elicited a multidimensional phobic trait that crossed observers, situations, and instruments. The results are discussed in terms of the implications for multivariant analysis of behavior change as well as for classification of psychopathological disorders of childhood. The thesis is advanced that instrument factors elicited by factor analysis in psychotherapy change studies are statistical artifacts and that a multiple discriminant analysis is a more appropriate technique for the study of change as well as for classification.