Beaten to the choice of instruments for classification or selection one has to take into account reliability, validity, representativeness, coverage and comprehensiveness on the symptom, syndromal and nosological level. Furthermore the strategies to handle heterogenity by way of atypy, hierarchical ordening, co-morbidity and life-time prevalence, multiaxial information, polydiagnostic approach, timeframing and thresholdlevels. Theory formulation underlying classification is missing and adequate research is failing. Computerized data without any theory are as prone to bias as clinical standardized procedures are; both without the possibility of falsifying. Standardized instruments differ importantly in their characteristics. There is no ideal instrument. The researcher is left to choose in accordance with his goals and directions given by his questions. It is naïve to try to classify disorders as well as patients into one categorical system.